Developmental plasticity in thermal tolerance is insufficient to compensate for rising temperatures: a meta-analysis
Supporting Information S2
Load packages and data
Load packages
pacman::p_load(tidyverse,
kableExtra,
DataExplorer,
viridis,
viridisLite,
maps,
rotl,
ape,
patchwork,
R.utils,
ggtree, # devtools::install_github("YuLab-SMU/ggtree")
ggtreeExtra, # devtools::install_github("xiangpin/ggtreeExtra")
phytools,
tidytree,
ggnewscale,
RColorBrewer,
metafor,
metaAidR,
orchaRd, # devtools::install_github("daniel1noble/orchaRd")
ggdist,
ggstatsplot,
here,
ggExtra,
emmeans,# remotes::install_github("rvlenth/emmeans")
MuMIn) Load data
d.raw <- read_csv("data/Raw_data.csv")
# Reformat ID variables as characters
d.raw$species_ID = as.character(d.raw$species_ID)
d.raw$study_ID = as.character(d.raw$study_ID)
d.raw$population_ID = as.character(d.raw$population_ID)
d.raw$family_ID = as.character(d.raw$family_ID)
d.raw$shared_trt_ID = as.character(d.raw$shared_trt_ID)
d.raw$cohort_ID = as.character(d.raw$cohort_ID)
d.raw$es_ID = as.character(d.raw$es_ID)
# Quick summary of the data set
kable(summary(d.raw), "html") %>%
kable_styling("striped", position = "left") %>%
scroll_box(width = "100%", height = "600px")| initials | es_ID | study_ID | species_ID | population_ID | family_ID | shared_trt_ID | cohort_ID | note_ID | data_source | data_url | fig_file_name | data_type | data_file_name | peer-reviewed | ref | title | pub_year | journal | thesis_chapter | doi | citation | phylum | class | order | family | genus | species | genus_species | age_maturity | ref_age_maturity | habitat | taxonomic_group | reproduction_mode | life_stage_manip | life_stage_tested | brought_common_temp | mobility_life_stage_manip | time_common_temp | common_temp | exp_design | origin_hatching | latitude | longitude | elevation | season | year | body_length | body_mass | age_tested | sex | housing_temp | incubation_independent | metric | endpoint | acc_temp_low | acc_temp_high | acc_temp_var | is_acc_temp_fluctuating | acc_duration | ramping | set_time | n_test_temp | n_replicates_per_temp | n_animals_per_replicate | humidity | oxygen | salinity | pH | photoperiod | gravidity | starved | minor_concerns | major_concerns | notes_moderators | mean_HT_low | sd_HT_low | n_HT_low | mean_HT_high | sd_HT_high | n_HT_high | error_type | notes_es | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Min. :1960 | Length:1131 | Min. :1.000 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Min. : 0.00821 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Min. : 0.125 | Min. : 6.0 | Length:1131 | Length:1131 | Min. :-41.87 | Min. :-128.23 | Min. : -14.0 | Length:1131 | Min. :1957 | Min. : 0.1782 | Min. : 0.000 | Min. : 0.00 | Length:1131 | Min. : 2.00 | Length:1131 | Length:1131 | Length:1131 | Min. : 0.50 | Min. : 5.00 | Min. : 0.0850 | Length:1131 | Min. : 0.9167 | Min. :0.00069 | Min. : 0.0028 | Min. : 2.000 | Min. : 1.000 | Min. : 1.00 | Min. :60.00 | Min. : 1.308 | Min. : 0.30 | Min. :6.800 | Min. : 8.00 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Length:1131 | Min. :14.19 | Min. : 0.02362 | Min. : 2.00 | Min. :15.38 | Min. : 0.0000 | Min. : 2.0 | Length:1131 | Length:1131 | |
| Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | 1st Qu.:1998 | Class :character | 1st Qu.:3.000 | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | 1st Qu.: 0.50000 | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | Class :character | 1st Qu.: 8.000 | 1st Qu.:14.9 | Class :character | Class :character | 1st Qu.: 26.83 | 1st Qu.: -96.91 | 1st Qu.: 0.0 | Class :character | 1st Qu.:2009 | 1st Qu.: 14.0000 | 1st Qu.: 0.310 | 1st Qu.: 5.00 | Class :character | 1st Qu.:17.00 | Class :character | Class :character | Class :character | 1st Qu.:13.00 | 1st Qu.:18.00 | 1st Qu.: 0.2000 | Class :character | 1st Qu.: 7.0000 | 1st Qu.:0.10000 | 1st Qu.: 1.0000 | 1st Qu.: 5.000 | 1st Qu.: 1.000 | 1st Qu.: 1.00 | 1st Qu.:65.00 | 1st Qu.: 5.500 | 1st Qu.:30.00 | 1st Qu.:7.550 | 1st Qu.:12.00 | Class :character | Class :character | Class :character | Class :character | Class :character | 1st Qu.:28.58 | 1st Qu.: 0.32097 | 1st Qu.: 8.00 | 1st Qu.:29.57 | 1st Qu.: 0.2901 | 1st Qu.: 7.0 | Class :character | Class :character | |
| Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Median :2012 | Mode :character | Median :4.000 | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Median : 1.79000 | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Median : 14.000 | Median :23.0 | Mode :character | Mode :character | Median : 37.26 | Median : -73.85 | Median : 30.0 | Mode :character | Median :2012 | Median : 48.0000 | Median : 1.900 | Median : 20.50 | Mode :character | Median :20.00 | Mode :character | Mode :character | Mode :character | Median :20.00 | Median :25.00 | Median : 0.3000 | Mode :character | Median : 14.0000 | Median :0.25000 | Median : 16.6667 | Median : 7.000 | Median : 3.000 | Median : 8.00 | Median :70.00 | Median : 6.600 | Median :30.00 | Median :7.963 | Median :12.00 | Mode :character | Mode :character | Mode :character | Mode :character | Mode :character | Median :35.00 | Median : 0.54000 | Median : 12.00 | Median :36.10 | Median : 0.5201 | Median : 11.0 | Mode :character | Mode :character | |
| NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | Mean :2006 | NA | Mean :3.707 | NA | NA | NA | NA | NA | NA | NA | NA | NA | Mean : 3.28359 | NA | NA | NA | NA | NA | NA | NA | NA | Mean : 33.233 | Mean :20.1 | NA | NA | Mean : 27.35 | Mean : -35.99 | Mean : 288.1 | NA | Mean :2008 | Mean : 57.9212 | Mean : 8.204 | Mean : 67.92 | NA | Mean :20.44 | NA | NA | NA | Mean :18.39 | Mean :23.16 | Mean : 0.5805 | NA | Mean : 21.9179 | Mean :0.38546 | Mean : 50.0755 | Mean : 7.761 | Mean : 3.453 | Mean : 13.07 | Mean :68.55 | Mean : 6.691 | Mean :29.81 | Mean :7.878 | Mean :12.89 | NA | NA | NA | NA | NA | Mean :34.21 | Mean : 0.82568 | Mean : 18.39 | Mean :35.01 | Mean : 0.7719 | Mean : 18.2 | NA | NA | |
| NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | 3rd Qu.:2018 | NA | 3rd Qu.:5.000 | NA | NA | NA | NA | NA | NA | NA | NA | NA | 3rd Qu.: 3.00000 | NA | NA | NA | NA | NA | NA | NA | NA | 3rd Qu.: 40.000 | 3rd Qu.:25.0 | NA | NA | 3rd Qu.: 49.85 | 3rd Qu.: 10.21 | 3rd Qu.: 285.0 | NA | 3rd Qu.:2016 | 3rd Qu.:100.0000 | 3rd Qu.:11.000 | 3rd Qu.: 60.00 | NA | 3rd Qu.:26.00 | NA | NA | NA | 3rd Qu.:24.00 | 3rd Qu.:28.00 | 3rd Qu.: 1.0000 | NA | 3rd Qu.: 30.0000 | 3rd Qu.:0.50000 | 3rd Qu.: 96.0000 | 3rd Qu.: 8.000 | 3rd Qu.: 5.000 | 3rd Qu.: 17.50 | 3rd Qu.:70.00 | 3rd Qu.: 7.300 | 3rd Qu.:35.00 | 3rd Qu.:8.130 | 3rd Qu.:14.00 | NA | NA | NA | NA | NA | 3rd Qu.:39.91 | 3rd Qu.: 0.99437 | 3rd Qu.: 20.00 | 3rd Qu.:40.73 | 3rd Qu.: 0.9305 | 3rd Qu.: 20.0 | NA | NA | |
| NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | Max. :2021 | NA | Max. :5.000 | NA | NA | NA | NA | NA | NA | NA | NA | NA | Max. :19.48528 | NA | NA | NA | NA | NA | NA | NA | NA | Max. :182.500 | Max. :36.0 | NA | NA | Max. :110.32 | Max. : 174.80 | Max. :3324.0 | NA | Max. :2019 | Max. :344.0000 | Max. :88.345 | Max. :758.00 | NA | Max. :30.00 | NA | NA | NA | Max. :37.00 | Max. :38.00 | Max. :10.0000 | NA | Max. :426.0000 | Max. :3.00000 | Max. :168.0000 | Max. :20.000 | Max. :24.000 | Max. :100.00 | Max. :76.00 | Max. :10.450 | Max. :40.00 | Max. :8.600 | Max. :16.00 | NA | NA | NA | NA | NA | Max. :53.90 | Max. :15.51098 | Max. :480.00 | Max. :58.70 | Max. :24.9888 | Max. :480.0 | NA | NA | |
| NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA’s :1056 | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA’s :146 | NA | NA | NA | NA | NA | NA | NA | NA | NA’s :926 | NA’s :908 | NA | NA | NA’s :823 | NA’s :823 | NA’s :828 | NA | NA’s :812 | NA’s :708 | NA’s :623 | NA’s :733 | NA | NA’s :384 | NA | NA | NA | NA | NA | NA’s :538 | NA | NA’s :305 | NA’s :272 | NA’s :858 | NA’s :880 | NA’s :878 | NA’s :901 | NA’s :1091 | NA’s :974 | NA’s :915 | NA’s :969 | NA’s :582 | NA | NA | NA | NA | NA | NA | NA’s :107 | NA’s :21 | NA | NA’s :113 | NA’s :21 | NA | NA |
# 1131 effect sizes 158 studies 147 species 251 populationsData processing
On 2022/01/13, we noticed few mistakes in the extracted data. We also had concerns about one study. Specifically, two different species_ID were assigned to Ischura elegans and Litopenaeus vannamei. We also found that Lampsilis abrupta, Lampsilis radiata and Lampsilis cariosa were mistakenly each assigned the same species_ID as Myxocyprinus asiaticus, Procypris rabaudi and Schizopygopsis younghusbandi, respectively.
The study from Buckley and Nufio (2014) Conservation Physiology, was also found not to be eligible to our inclusion criteria. The main author of the study indicated to us that “The acclimation to 25C was short term (1-2 hours) immediately before CTmax measurements. Animals were raised at either 24 or 27C until adulthood (and CTmax measurements)” which suggests an overlap between juvenile and adult acclimation. In additon, the re-acclimation period turned out to be short to be considered relevant in our study.
On 2022/01/21, we also realised that the equation 6 (imputation of missing standard deviations) is not valid to impute missing standard errors. Therefore, we had to exclude all observations for which the standard error was missing (studies from: Rosa_et_al_2014; Campbell_et_al_2020; Blaxter 1960; Hokanson & Koenst 1986; Brown_1969; Edsall_and_Colby_1970; Reynolds_1976; Threader_and_Houston_1983). These observations were all LT50 measurements, for which sample standard deviation does not make particular sense because the measure of dispersion (standard error) is inferred from regression-based interpolations.
There was also one effect size (es_ID 1009 from Re_et_al_2006) for which the standard deviation was taken as “0”. This value was replaced as “NA”.
# Identify different species IDs used from the same species
d.raw$species_ID[d.raw$genus_species == "Ischnura elegans"]## [1] "15" "15" "15" "15" "15" "15" "15" "15" "15" "15" "15" "16" "15" "15" "15"
## [16] "15"d.raw$species_ID[d.raw$genus_species == "Litopenaeus vannamei"]## [1] "79" "79" "79" "79" "79" "79" "79" "79" "79" "79" "79" "79" "50" "50" "50"
## [16] "50" "50" "50"# Identify different species with the same species ID
d.raw$genus_species[d.raw$species_ID == "109"]## [1] "Myxocyprinus asiaticus" "Lampsilis abrupta" "Lampsilis abrupta"
## [4] "Lampsilis abrupta"d.raw$genus_species[d.raw$species_ID == "110"]## [1] "Procypris rabaudi" "Lampsilis radiata"d.raw$genus_species[d.raw$species_ID == "111"]## [1] "Lampsilis cariosa" "Lampsilis cariosa"
## [3] "Schizopygopsis younghusbandi" "Schizopygopsis younghusbandi"
## [5] "Schizopygopsis younghusbandi" "Schizopygopsis younghusbandi"# Correct mistakes
d.raw$species_ID[d.raw$genus_species == "Ischnura elegans"] <- "15" # Same species_ID for this species
d.raw$species_ID[d.raw$genus_species == "Litopenaeus vannamei"] <- "50" # Same species_ID for this species
d.raw$species_ID[d.raw$genus_species == "Lampsilis abrupta"] <- "148" # Use number not used previously
d.raw$species_ID[d.raw$genus_species == "Lampsilis radiata"] <- "149" # Use number not used previously
d.raw$species_ID[d.raw$genus_species == "Lampsilis cariosa"] <- "150" # Use number not used previously
# Exclude the study from Buckley and Nufio (2014)
d.raw <- filter(d.raw, ref != "Buckley_and_Nufio_2014")
d.raw$sd_HT_high[d.raw$sd_HT_high == "0"] <- NA # Replace the null SD value by NA
# Identify effect sizes where the measure of dispersion is standard error, and
# sample size is missing
d.raw$ref[d.raw$error_type == "se" & is.na(d.raw$sd_HT_high) == "TRUE"]## [1] "Rosa_et_al_2014" "Campbell_et_al_2020"
## [3] "Blaxter_1960" "Blaxter_1960"
## [5] "Blaxter_1960" "Brown_1969"
## [7] "Brown_1969" "Brown_1969"
## [9] "Brown_1969" "Brown_1969"
## [11] "Brown_1969" "Brown_1969"
## [13] "Brown_1969" "Brown_1969"
## [15] "Brown_1969" "Brown_1969"
## [17] "Brown_1969" "Brown_1969"
## [19] "Brown_1969" "Brown_1969"
## [21] "Edsall_and_Colby_1970" "Edsall_and_Colby_1970"
## [23] "Edsall_and_Colby_1970" "Edsall_and_Colby_1970"
## [25] "Hokanson_and_Koenst_1986" "Hokanson_and_Koenst_1986"
## [27] "Reynolds_1976" "Reynolds_1976"
## [29] "Threader_and_Houston_1983" "Threader_and_Houston_1983"
## [31] "Threader_and_Houston_1983" "Threader_and_Houston_1983"
## [33] "Threader_and_Houston_1983" "Threader_and_Houston_1983"
## [35] "Threader_and_Houston_1983" "Threader_and_Houston_1983"
## [37] "Threader_and_Houston_1983" "Threader_and_Houston_1983"
## [39] "Threader_and_Houston_1983" "Threader_and_Houston_1983"d.raw <- mutate(d.raw, exclude = ifelse(error_type == "se" & is.na(sd_HT_high) ==
"TRUE" | error_type == "se" & is.na(sd_HT_low) == "TRUE", "yes", "no")) # If the standard error is missing for either the low or high acclimation temperature, indicate exclude = 'yes', otherwise 'no'
d.raw <- filter(d.raw, exclude != "yes") # Exclude all observations where the standard error is missingCalculate effect sizes
Impute missing standard deviations
Missing standard deviation was estimated using formula 13.1 (p. 199) of the Handbook of Meta-anlysis in Ecology and Evolution (Koricheva, Gurevitch and Mengersen, 2013)
Because treatments are repeated in stepwise multiple comparisons (i.e., the same data is used twice to calculate different effect sizes), we made sure the imputed standard deviations (or standard error) were the same for the repeated data.
First, only keep data with SD (because we can only impute SD), then do the within/between study means, then impute.
# First, only take observations with standard deviations because they are the only ones we can impute
d.sd<-filter(d.raw, error_type=="sd")
# Check assumption that SD/mean ratio is relatively constant between studies
d.sd %>% group_by(study_ID) %>% # Group each study together
mutate(within_study_mean_low = mean(mean_HT_low, na.rm = T), # Calculate the within study heat tolerance at the lowest acclimation temperature
within_study_mean_high = mean(mean_HT_high, na.rm = T), # Calculate the within study heat tolerance at the highest acclimation temperature
within_study_sd_low = mean(sd_HT_low, na.rm=T), # Calculate the within study standard deviation at the lowest acclimation temperature
within_study_sd_high =mean(sd_HT_high, na.rm=T), # Calculate the within study standard deviation at the lowest acclimation temperature
sd_mean_ratio_low = within_study_sd_low / within_study_mean_low, # Calculate the mean to SD ratio for lowest acclimation temperatures
sd_mean_ratio_high = within_study_sd_high / within_study_mean_high) %>% # Calculate the mean to SD ratio for highest acclimation temperatures
ggplot() + geom_histogram(aes(x=sd_mean_ratio_low), fill="red", alpha=0.2) + # Plot the results
geom_histogram(aes(x=sd_mean_ratio_high), fill="blue", alpha=0.2)
# Indicate which treatment is re-used in >1 comparisons
d.trt <- group_by(d.sd,shared_trt_ID) %>% summarise(n_trt=n())
d.sd <-left_join(d.sd, d.trt, by="shared_trt_ID")
# Indicate which cohort is repeated >1 times to calculate sampling variances of non-independent observations differently.
d.cohort <- group_by(d.sd,cohort_ID) %>% summarise(n_cohort=n())
d.sd<-left_join(d.sd, d.cohort, by="cohort_ID")
# Calculate within and between study heat tolerance and standard deviation
d.sd <- d.sd %>%
group_by(study_ID) %>% # Group by study
mutate(within_study_mean_low = mean(mean_HT_low, na.rm = T), # Calculate within study heat tolerance at the lowest acclimation temperature
within_study_mean_high = mean(mean_HT_high, na.rm = T), # Calculate within study heat tolerance at the highest acclimation temperature
within_study_sd_low = mean(sd_HT_low, na.rm=T), # Calculate within study standard deviation at the lowest acclimation temperature
within_study_sd_high = mean(sd_HT_high, na.rm=T)) %>% # Calculate within study standard deviation at the highest acclimation temperature
ungroup() %>% # Ungroup observations to calculate between study heat tolerance and SD
mutate(
between_study_mean_low=mean(within_study_mean_low, na.rm=T), # Calculate between study heat tolerance at the lowest acclimation temperature
between_study_mean_high=mean(within_study_mean_high, na.rm=T), # Calculate between study heat tolerance at the highest acclimation temperature
between_study_sd_low=mean(within_study_sd_low, na.rm=T), # Calculate between study standard deviation at the lowest acclimation temperature
between_study_sd_high=mean(within_study_sd_high, na.rm=T) # Calculate between study standard deviation at the highest acclimation temperature
)
d.sd$sd_HT_high[is.na(d.sd$sd_HT_high)]<-"missing" # Problem detecting NA so replaced NA with "missing"
d.sd$sd_HT_low[is.na(d.sd$sd_HT_low)]<-"missing"
d.sd <- d.sd %>% mutate(imputed=ifelse(sd_HT_low=="missing"|sd_HT_high=="missing", "yes", "no")) # Add a column `imputed`, indicated whether or not the standard deviation is missing and need to be imputed
# Impute missing standard deviations
d.sd <- mutate(d.sd,
imputed_sd_low = ifelse(sd_HT_low=="missing"&n_trt!="1",
(within_study_mean_high*(between_study_sd_high/between_study_mean_high)),
ifelse(sd_HT_low=="missing"&n_trt=="1",
(within_study_mean_low*(between_study_sd_low/between_study_mean_low)), sd_HT_low)), # If the treatments are repeated, take the sd of the high acclimation group (because we made stepwise comparisons), otherwise estimate sd normally.
imputed_sd_high = ifelse(sd_HT_high=="missing",
(within_study_mean_high*(between_study_sd_high/between_study_mean_high)), sd_HT_high))
d.sd$imputed_sd_low = as.numeric(d.sd$imputed_sd_low) # Force sd to be considered as numeric values
d.sd$imputed_sd_high = as.numeric(d.sd$imputed_sd_high) # Force sd to be considered as numeric values
d.sd$sd_HT_low = as.numeric(d.sd$sd_HT_low) # Force sd to be considered as numeric values
d.sd$sd_HT_high = as.numeric(d.sd$sd_HT_high) # Force sd to be considered as numeric values
d.se<-filter(d.raw, error_type=="se") # Filter the data where the measure of dispersion is SE
d.se$imputed<- "no" # Assign the value "no" for the column "imputed"
# Indicate which treatment is re-used in >1 comparisons
d.trt.se <- group_by(d.se,shared_trt_ID) %>% summarise(n_trt=n())
d.se <-left_join(d.se, d.trt.se, by="shared_trt_ID") # Join this information to d.se
# Indicate which cohort is repeated >1 times to calculate sampling variances of non-independent observations differently.
d.cohort.se <- group_by(d.se,cohort_ID) %>% summarise(n_cohort=n())
d.se<-left_join(d.se, d.cohort.se, by="cohort_ID") # Join this information to d.se
d<-full_join(d.sd, d.se) # Join the data with imputed standard deviations
d$es_ID=as.numeric(d$es_ID) # Convert effect size ID to numeric to order the data
d<-arrange(d, es_ID) # Order the data by effect size ID
d$es_ID=as.factor(d$es_ID) # Convert effect size ID back to a factor
# General diagnostic plots
ggplot(d, aes(x=study_ID, y=log10(imputed_sd_low), col=imputed, shape=metric), show.legend=F)+geom_point(size=2, alpha=0.5)
ggplot(d, aes(x=study_ID, y=log10(imputed_sd_high), col=imputed, shape=metric), show.legend=F)+geom_point(size=2, alpha=0.5)
Calculate effect sizes and sampling variances
# Calculate effect sizes (formulas are presented in the manuscript)
d <- d %>% mutate(dARR = ((mean_HT_high-mean_HT_low)/(acc_temp_high-acc_temp_low)),
Var_dARR = ifelse(error_type=="sd"&n_cohort=="1",
((1/(acc_temp_high - acc_temp_low))^2*(imputed_sd_low^2/n_HT_low+imputed_sd_high^2/n_HT_high)), # If SD and independent, use this formula
ifelse(error_type=="sd"&n_cohort!="1", # If SD and non-independent, use this formula
((1/(acc_temp_high - acc_temp_low))^2*((imputed_sd_low^2 + imputed_sd_high^2 - 2*0.5*imputed_sd_low*imputed_sd_high)/(n_HT_low + n_HT_high))),
ifelse(error_type=="se"&n_cohort=="1", # If SE and independent, use this formula
((1/(acc_temp_high - acc_temp_low))^2*(sd_HT_low^2+sd_HT_high^2)),
((1/(acc_temp_high - acc_temp_low))^2*(sd_HT_low^2+ sd_HT_high^2 - 2*0.5*sd_HT_low*sd_HT_high))))), # Otherwise (if SE and non-independent, use this last formula)
precision =(1/sqrt(Var_dARR))) # Calculate precision (1/SE)
summary(d$dARR) # Summary of the effect sizes## Min. 1st Qu. Median Mean 3rd Qu. Max.
## -2.41514 0.02227 0.14704 0.16981 0.30649 2.70000summary(d$Var_dARR) # Summary of the sampling variance of dARR; seems fine except for the maximum## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.00001 0.00074 0.00319 0.17200 0.01446 34.73580summary(d$precision) # Summary of the precision; seems fine except for the maximum## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.1697 8.3163 17.7139 29.0091 36.7072 398.5101ggplot(d, aes(x=study_ID, y=log10(Var_dARR), col=imputed, shape=metric), show.legend=F)+geom_point(size=2, alpha=0.5) 
ggplot(d, aes(x=study_ID, y=precision, col=imputed, shape=metric), show.legend=F)+geom_point(size=2, alpha=0.5)
Merge designs E and F into the same design and rename design G
Originally, the design presented as design E in Figure 2 of the manuscript was separated into two designs “E” and “F”, with the current design “F” being classified as “G”. The difference between the former designs “E” and “F” was on the life stage tested (i.e., former design “E” only comprised data from animals tested at the juveniles stage, whereas former design “F” only comprised data from animals tested right after the adult transition; although both designs shared similar acclimation periods)
Because designs E and F are very similar, and the sample size for design F was very low (2 studies), we decided to merge these designs. Design G was renamed to design “F”.
d$exp_design <- str_replace(d$exp_design, "F", "E") # Merge designs 'E' and 'F'
d$exp_design <- str_replace(d$exp_design, "G", "F") # Rename 'G' as 'F'Data exploration
Check for extreme values and/or studies
Here, we checked the mean, minimum, and maximum values of dARR and Var_dARR for each study to see if the extreme values are particularly condensed in certain studies.
We also plotted the distribution of dARR to see if the response variable is skewed.
kable(d %>%
group_by(ref) %>% # Group by study
summarise(mean_dARR = mean(dARR),
var_dARR = mean(Var_dARR),
prec = mean(precision),
max_dARR = max(dARR),
min_dARR = min(dARR),
n = n())) %>% # Calculate descriptive statistics
kable_styling("striped", position = "left") %>% # Stylize the table
scroll_box(width = "100%", height = "500px")| ref | mean_dARR | var_dARR | prec | max_dARR | min_dARR | n |
|---|---|---|---|---|---|---|
| Abayarathna_et_al_2019 | 0.0422826 | 0.0001814 | 75.3091248 | 0.0540130 | 0.0115391 | 5 |
| Akhtar_et_al_2013 | 0.4549679 | 0.0084695 | 13.0618648 | 0.6982367 | 0.2525533 | 8 |
| Alford_2010 | 0.0644712 | 0.0001677 | 102.4163159 | 0.1886680 | -0.0404940 | 36 |
| Alford_et_al_2012 | 0.0194444 | 0.0005000 | 53.0330086 | 0.1200000 | -0.0300000 | 18 |
| Archambault_et_al_2014a | 0.2500000 | 0.0312370 | 5.8724124 | 0.2800000 | 0.2200000 | 2 |
| Archambault_et_al_2014b | 0.0485714 | 0.0153061 | 8.3890929 | 0.5400000 | -0.2600000 | 7 |
| Ashaf-Ud-Doulah_et_al_2020 | -0.3333333 | 0.0061033 | 12.8904219 | -0.1666667 | -0.5000000 | 2 |
| Asheim_et_al_2020 | 0.2452538 | 0.0000084 | 352.9905278 | 0.3515975 | 0.1389100 | 2 |
| Azra_et_al_2018 | 0.3176913 | 0.0048178 | 14.4070170 | 0.4858800 | 0.2616275 | 4 |
| Bai_et_al_2019 | -0.0979187 | 0.0047536 | 14.9844055 | 0.0706650 | -0.3286250 | 4 |
| Baroudy_and_Elliott_1994 | 0.0764341 | 0.0056489 | 21.4770868 | 0.4680000 | -0.3000000 | 43 |
| Becker_and_Genoway_1979 | 0.2663292 | 0.0001102 | 102.4939799 | 0.3810130 | 0.1873420 | 10 |
| Benedict_et_al_1991 | 0.1135920 | 0.0004995 | 44.7442857 | 0.1135920 | 0.1135920 | 1 |
| Billman_et_al_2008 | 0.4135680 | 2.7801338 | 13.6085427 | 0.8333333 | -0.0072440 | 9 |
| Bishai_1960 | 0.2787259 | 0.0350696 | 6.7536342 | 0.7180689 | 0.0773267 | 6 |
| Blair_and_Glover_2019 | -0.0547942 | 0.0003733 | 51.7583539 | -0.0547942 | -0.0547942 | 1 |
| Bowden_et_al_2018 | 0.3560112 | 0.0292445 | 5.8630533 | 0.4881225 | 0.1866100 | 4 |
| Britton_2005 | 0.0697917 | 0.0005755 | 52.3497327 | 0.5500000 | -0.1666667 | 16 |
| Bugg_et_al_2020 | 0.3592606 | 0.0025333 | 21.8863872 | 0.5742775 | 0.1861925 | 4 |
| Cai_and_Chen_2005 | 0.0455578 | 0.0169639 | 11.8866227 | 0.2866667 | -0.7535900 | 18 |
| Carbonell_and_Stoks_2020 | 0.1721478 | 0.0087696 | 13.4077909 | 0.4118775 | 0.0095800 | 9 |
| Chaterjee_et_al_2004 | 0.2650000 | 0.0005862 | 46.5140866 | 0.4600000 | 0.1200000 | 8 |
| Chen_and_Chen_1999 | 0.1843035 | 0.0023786 | 20.6360961 | 0.5400000 | 0.0489440 | 12 |
| Chen_et_al_2013 | -0.2500000 | 0.0039815 | 55.9404777 | 0.0750000 | -0.5500000 | 22 |
| Cheung_2019 | -0.1015827 | 0.0308685 | 9.2756558 | 0.6157962 | -0.7636286 | 43 |
| Chidawanyika_and_Terblanche_2010 | 0.1146790 | 0.0002147 | 68.2552997 | 0.2293580 | 0.0000000 | 2 |
| Cook_et_al_2006 | 0.3565646 | 0.0003226 | 68.8543839 | 0.5600000 | 0.1166667 | 7 |
| Currie_et_al_1998 | 0.2933333 | 0.0012290 | 30.7886049 | 0.4600000 | 0.1400000 | 6 |
| Dang_et_al_2019 | 0.0853788 | 0.0003908 | 51.3033640 | 0.1097410 | 0.0658450 | 4 |
| Das_et_al_2004 | 0.2032407 | 0.0082969 | 28.8623767 | 0.6300000 | 0.0333333 | 18 |
| Das_et_al_2005 | 0.2831111 | 0.0010057 | 38.9740108 | 0.4960000 | 0.0833333 | 3 |
| Dayanada_et_al_2017 | -0.3985789 | 0.0037211 | 16.6430034 | -0.3851684 | -0.4119895 | 2 |
| de-Beek_et_al_2017 | 0.0826088 | 0.0023948 | 20.4422841 | 0.0852850 | 0.0799325 | 2 |
| de-Beek_et_al_2018a | 0.0698550 | 0.0017264 | 24.0675422 | 0.0698550 | 0.0698550 | 1 |
| de-Beek_et_al_2018b | 0.0992363 | 0.0020452 | 22.1610679 | 0.1547075 | 0.0447825 | 4 |
| Debnath_et_al_2006 | 0.1843750 | 0.0003500 | 87.9325151 | 0.3500000 | 0.0450000 | 4 |
| Del-Rio_et_al_2019 | -0.0168266 | 0.0045805 | 19.0698223 | 0.2670950 | -0.5235050 | 8 |
| Delorme-Juri_2017 | 0.1444700 | 15.4177909 | 1.3235837 | 1.1854700 | -0.8413000 | 6 |
| Deslauriers_et_al_2016 | 0.2104328 | 0.0043057 | 19.4854456 | 0.2791660 | 0.1495499 | 3 |
| Diaz-Herrera_et_al_1998 | 0.4375000 | 0.0205716 | 11.5153366 | 0.7666667 | 0.1666667 | 8 |
| Donelson_et_al_2015 | 0.0592972 | 0.0014255 | 26.6957898 | 0.2898975 | -0.1713031 | 2 |
| Dulger_et_al_2012 | 0.2710000 | 0.0011877 | 30.1827669 | 0.3220000 | 0.2200000 | 2 |
| Ebel_et_al_1971 | -0.0357366 | 3.3237193 | 0.6891243 | 0.7739900 | -0.7083880 | 10 |
| Elliott_1991 | 0.1255392 | 0.0201177 | 9.6820608 | 0.5781991 | -0.5781990 | 80 |
| Elliott_and_Klemetsen_2002 | 0.1438095 | 0.0028514 | 22.9464408 | 0.3479953 | 0.0253548 | 8 |
| Esquer-Mendez_et_al_2010 | 0.1041667 | 0.0126217 | 20.9382859 | 0.3125000 | 0.0000000 | 3 |
| Faleiro_et_al_2016 | 0.4285714 | 0.6356765 | 1.2542437 | 0.4285714 | 0.4285714 | 1 |
| Fields_et_al_1987 | 0.4760417 | 0.0014669 | 28.9182905 | 0.5750000 | 0.3625000 | 12 |
| Floyd_1983 | 0.1980737 | 0.0007774 | 63.7249643 | 0.3875354 | -0.1314448 | 20 |
| Floyd_1985 | 0.1647468 | 0.0005790 | 90.0607240 | 0.2375000 | 0.0692641 | 13 |
| Gervais_et_al_2020 | 0.5077778 | 0.0038111 | 17.3971720 | 0.6466667 | 0.3700000 | 3 |
| Gibson_et_al_2015 | -0.0908345 | 0.0335600 | 14.3738981 | 0.0412700 | -0.2187500 | 4 |
| Gomez-Isaza_et_al_2020 | 0.2348710 | 0.0010724 | 30.5372505 | 0.2348710 | 0.2348710 | 1 |
| Gray_2013 | 0.0386719 | 0.0005119 | 44.4515464 | 0.0719863 | 0.0053575 | 2 |
| Gunderson_et_al_2020 | 0.0357143 | 0.0842502 | 3.6202384 | 0.6428571 | -0.5000000 | 4 |
| He_et_al_2014 | 0.1681250 | 0.0000594 | 154.0989112 | 0.2550000 | 0.0500000 | 4 |
| Healy_et_al_2019 | 0.1048000 | 0.0009741 | 33.9768257 | 0.2412500 | -0.0425000 | 10 |
| Illing_et_al_2020 | 0.2311021 | 0.1328116 | 7.4400875 | 1.6074000 | -0.3640778 | 27 |
| Ineno_et_al_2005 | 0.0902312 | 0.0208194 | 42.7221821 | 1.0714375 | -1.1207650 | 13 |
| Kelly_et_al_2014 | 0.2495770 | 0.0041549 | 18.0829434 | 0.5074299 | 0.0269977 | 12 |
| Kerfoot_2012 | 0.0000000 | 0.0147221 | 8.2416799 | 0.1600000 | -0.1600000 | 2 |
| Kern_et_al_2015 | 0.2465955 | 0.0003993 | 50.0752545 | 0.2817210 | 0.2114700 | 2 |
| Kingsolver_et_al_2016 | 0.2283242 | 0.0193378 | 9.4239779 | 0.4798000 | -0.2197000 | 4 |
| Kir_2020 | 0.2560000 | 0.0123074 | 18.4841046 | 0.7460000 | 0.0720000 | 6 |
| Kir_and_Demirci_2018 | 0.3096667 | 0.0120644 | 18.6234218 | 0.5780000 | 0.1800000 | 6 |
| Kir_et_al_2017 | 0.2300000 | 0.0007922 | 48.2226547 | 0.2675000 | 0.1650000 | 3 |
| Komoroske_et_al_2014 | 0.2434211 | 0.0022442 | 21.3182700 | 0.7368421 | -0.2500000 | 2 |
| Kumlu_et_al_2010 | 0.4203333 | 0.0015146 | 34.5051064 | 0.5420000 | 0.3300000 | 6 |
| Larios-Soriano_et_al_2019 | 0.3785242 | 0.0032903 | 18.4470864 | 0.4742933 | 0.3101167 | 4 |
| Leon-Palomino_et_al_2017 | 0.4899107 | 0.0121333 | 11.3917243 | 1.0609733 | -0.0811520 | 2 |
| Li_et_al_2015 | 0.1781333 | 0.0087457 | 15.3922944 | 0.4950000 | -0.0780000 | 5 |
| Linton_et_al_1998 | 0.2111598 | 0.0144344 | 8.7693821 | 0.2482579 | 0.1572316 | 3 |
| Llewelyn_et_al_2018 | 0.0177900 | 0.0003837 | 54.0875299 | 0.0753533 | -0.0258900 | 4 |
| Lohr_et_al_1996 | 0.1847002 | 0.0102181 | 21.6696135 | 0.2763158 | 0.0963850 | 4 |
| Lu_et_al_2020 | 0.3100000 | 0.1027122 | 3.1202464 | 0.3100000 | 0.3100000 | 1 |
| Ma_and_Ma_2012 | 0.0677063 | 0.0013982 | 27.0361051 | 0.1603567 | -0.0249440 | 2 |
| Manriquez_et_al_2019 | 0.1229922 | 0.0169003 | 13.9849521 | 0.1976285 | 0.0483559 | 2 |
| Mascaro_et_al_2016 | 0.6416667 | 0.0004225 | 48.6530912 | 0.6416667 | 0.6416667 | 1 |
| Mascaro_et_al_2017 | 0.3916667 | 0.0015799 | 25.1588361 | 0.3916667 | 0.3916667 | 1 |
| Mascaro_et_al_2018 | 0.2583333 | 0.0009954 | 31.6962326 | 0.2583333 | 0.2583333 | 1 |
| McCauley_1963 | 0.0310620 | 0.0246195 | 6.3732458 | 0.0310620 | 0.0310620 | 1 |
| McDermid_et_al_2013 | 0.3145695 | 0.0004557 | 47.2930566 | 0.3791380 | 0.2781460 | 4 |
| McDonnell_et_al_2019 | 0.3054676 | 0.0005401 | 43.0861315 | 0.4515736 | -0.0977970 | 4 |
| Medina-Romo_et_al_2018 | 0.3580833 | 0.0171857 | 7.6280912 | 0.6319133 | 0.1053167 | 4 |
| Meng_et_al_2009 | 0.0900000 | 0.0006508 | 39.2000063 | 0.0900000 | 0.0900000 | 1 |
| Mitchell_et_al_1993 | 0.1072580 | 0.0002899 | 61.8369460 | 0.3322580 | -0.1935480 | 4 |
| Moyano_et_al_2017 | 0.5984217 | 0.0404443 | 6.6869073 | 1.1739923 | 0.2091346 | 7 |
| Mueller_et_al_2019 | 0.0086520 | 0.0102290 | 10.8981737 | 0.3987168 | -0.2554760 | 6 |
| Munoz_et_al_2018 | 0.2504176 | 0.0019224 | 24.2865463 | 0.3494455 | 0.0910614 | 4 |
| Mutamiswa_et_al_2018a | 0.0234410 | 0.0011717 | 29.2183745 | 0.1253600 | -0.0445360 | 6 |
| Mutamiswa_et_al_2018b | 0.0499230 | 0.0010794 | 30.4388650 | 0.1248080 | -0.0249620 | 2 |
| Myrick_1998 | 0.1443156 | 0.1622640 | 17.8326161 | 0.2692308 | 0.0126697 | 8 |
| Myrick_and_Cech_2000 | 0.1333333 | 0.0093889 | 10.3203137 | 0.1333333 | 0.1333333 | 1 |
| Myrick_and_Cech_2002 | 0.3333333 | 0.0080222 | 11.1648439 | 0.3333333 | 0.3333333 | 1 |
| Myrick_and_Cech_2003 | 0.2175000 | 0.0067300 | 14.8718114 | 0.2666667 | 0.1750000 | 6 |
| Myrick_and_Cech_2005 | 0.2632875 | 0.0114797 | 9.5079355 | 0.3001975 | 0.2263775 | 2 |
| Noyola_et_al_2013 | 0.3916667 | 0.0224516 | 7.1393644 | 0.5250000 | 0.2250000 | 3 |
| Opuszynski_et_al_1989 | 0.3451461 | 0.0327028 | 5.5582393 | 0.5683000 | 0.1900767 | 3 |
| Orille_et_al_2020 | 0.2517306 | 0.0003014 | 57.6012050 | 0.2517306 | 0.2517306 | 1 |
| Oyamaguchi_et_al_2018 | 0.2316729 | 0.0229617 | 10.1583883 | 0.4608950 | 0.0942571 | 3 |
| Pandolfo_et_al_2010a | -0.1628571 | 0.2628369 | 2.3280499 | 0.5200000 | -1.1400000 | 7 |
| Pandolfo_et_al_2010b | -0.6866667 | 3.0780230 | 5.7405113 | -0.1400000 | -1.5000000 | 3 |
| Peng_et_al_2014 | 0.5122810 | 0.0012702 | 28.2565280 | 0.6363620 | 0.2747940 | 4 |
| Pereira_et_al_2017 | 0.1857273 | 0.0009366 | 34.1994500 | 0.2682000 | 0.0792000 | 11 |
| Perez_et_al_2003 | 0.4750000 | 0.0075652 | 11.4971338 | 0.4750000 | 0.4750000 | 1 |
| Perez_et_al_2004 | 0.6750000 | 0.0013015 | 27.7185858 | 0.6750000 | 0.6750000 | 1 |
| Perez_et_al_2006 | 0.0500000 | 0.0061823 | 12.7181594 | 0.0500000 | 0.0500000 | 1 |
| Perez_et_al_2007 | 0.1750000 | 0.0005694 | 41.9065627 | 0.1750000 | 0.1750000 | 1 |
| Perez_et_al_2010 | 0.7500000 | 0.0010575 | 30.7510100 | 0.7500000 | 0.7500000 | 1 |
| Perez_et_al_2012 | 0.0500000 | 0.0013015 | 27.7185858 | 0.0500000 | 0.0500000 | 1 |
| Pimentel_et_al_2014 | 0.3153800 | 0.0027002 | 20.3116342 | 0.4723500 | 0.1497700 | 4 |
| Piyaphongkul_et_al_2014 | 0.1970238 | 0.0211766 | 15.3007733 | 0.4714286 | 0.0125000 | 6 |
| Piyaphongkul_et_al_2018 | 0.2050000 | 0.0008500 | 35.0419485 | 0.2400000 | 0.1700000 | 2 |
| Porter_2016 | 0.2750000 | 0.1053125 | 3.0814849 | 0.2750000 | 0.2750000 | 1 |
| Porter_2019 | -0.0250000 | 0.0428125 | 4.8329767 | -0.0250000 | -0.0250000 | 1 |
| Porter_2021 | 0.6500000 | 0.0365625 | 5.2297636 | 0.6500000 | 0.6500000 | 1 |
| Procarione_and_King_1993 | 0.3980294 | 0.0102526 | 10.2374962 | 0.5555556 | 0.3076923 | 4 |
| Re_et_al_2006 | 0.4767143 | 0.0240879 | 8.2254399 | 1.2200000 | -0.0900000 | 35 |
| Re_et_al_2012 | 0.4250972 | 0.0020867 | 22.9887521 | 0.7766800 | 0.1290733 | 12 |
| Reyes_et_al_2011 | 0.2935600 | 0.0032982 | 17.5455199 | 0.3490233 | 0.1931133 | 4 |
| Rodgers_et_al_2019 | 0.3800000 | 0.0039071 | 16.1261792 | 0.4614286 | 0.2985714 | 2 |
| Rosa_et_al_2014 | 0.2467058 | 0.0344412 | 9.5418595 | 0.8278632 | -0.0604048 | 7 |
| Ruthsatz_et_al_2018 | 0.0333333 | 0.0152778 | 9.6199463 | 2.7000000 | -1.4000000 | 4 |
| Ruthsatz_et_al_2020 | 0.4927083 | 0.0081047 | 21.6126127 | 1.2266667 | -0.3933333 | 4 |
| Sakurai_et_al_2021 | 0.1557799 | 0.0005417 | 44.9398975 | 0.2081410 | 0.1034188 | 2 |
| Salachan_and_Sorensen_2017 | 0.0662500 | 0.0014432 | 27.7374492 | 0.1926675 | -0.0515000 | 4 |
| Salinas_et_al_2019 | 0.6607434 | 0.0639706 | 11.8925715 | 1.0468267 | 0.3064920 | 3 |
| Sasaki_and_Dam_2019 | 0.7621810 | 0.0251199 | 7.2983358 | 1.5226225 | 0.3175750 | 10 |
| Sasaki_and_Dam_2020 | 0.3297967 | 0.0029592 | 19.7639473 | 0.5723633 | -0.1087100 | 5 |
| Sasaki_et_al_2019 | 0.1908150 | 0.0162258 | 8.0619543 | 0.3579550 | 0.0359850 | 4 |
| Sasmita_et_al_2019 | 0.3458333 | 0.0895791 | 4.6069208 | 1.0600000 | 0.1066667 | 6 |
| Schaefer_and_Ryan_2006 | 0.0748408 | 0.0027237 | 19.1827253 | 0.1110675 | 0.0322450 | 3 |
| Shinner_et_al_2020 | 0.0058965 | 0.0003282 | 55.2042580 | 0.0168204 | -0.0050275 | 2 |
| Shrode_1975 | 0.0102783 | 0.0117458 | 9.2269509 | 0.1767225 | -0.2011500 | 13 |
| Simon_et_al_2015 | 0.1287012 | 0.0010855 | 44.2932737 | 0.2966670 | 0.0115390 | 15 |
| Slotsbo_et_al_2016 | 0.0719253 | 0.0015568 | 53.4677447 | 0.1457140 | -0.2067710 | 32 |
| Spinks_et_al_2019 | -0.0296650 | 0.0014100 | 29.8094881 | 0.0202000 | -0.1284100 | 8 |
| Stitt_2012 | 0.1727788 | 0.0008875 | 34.8665097 | 0.3130041 | 0.0692397 | 9 |
| Stoler_2012 | -0.1868946 | 0.0450393 | 10.5843418 | 0.7115385 | -1.2500000 | 3 |
| Tatum_2018 | 0.1928571 | 0.0063454 | 13.2625824 | 0.2285714 | 0.1500000 | 3 |
| Terblanche_and_Chown_2006 | -0.0028871 | 0.0014489 | 27.8831569 | 0.0279070 | -0.0336812 | 2 |
| Troia_et_al_2015 | 0.2790809 | 0.5391901 | 3.1993095 | 2.2348500 | -0.7703100 | 17 |
| Tsuchida_and_Setoguma_1997 | 0.2208561 | 0.0012408 | 42.0244152 | 0.3177400 | 0.1255080 | 9 |
| Underwood_et_al_2012 | 0.1937075 | 0.0025186 | 23.1729833 | 0.3333333 | 0.1224490 | 5 |
| Uriarte_et_al_2018 | 0.1914155 | 0.0148079 | 8.2177450 | 1.1422950 | -2.4151450 | 22 |
| Vanvelk_et_al_2021 | 0.1032731 | 0.0019630 | 23.1124143 | 0.2226817 | -0.0636233 | 6 |
| Wagner_et_al_2001 | 0.2261674 | 0.0006738 | 49.5270258 | 0.4090909 | -0.0195682 | 11 |
| Walsh_et_al_1998 | 0.2230000 | 0.0091460 | 10.4564533 | 0.2230000 | 0.2230000 | 1 |
| Wang_2014 | 0.3933333 | 0.0032844 | 18.0464599 | 0.7200000 | 0.1800000 | 3 |
| Wang_et_al_2013 | 0.3503160 | 0.0131409 | 10.1368526 | 0.4993660 | 0.2012660 | 2 |
| Warriner_et_al_2020 | 0.2000000 | 0.0006444 | 39.3919299 | 0.2000000 | 0.2000000 | 1 |
| White_and_Wahl_2020 | 0.6053272 | 0.0030467 | 20.0036686 | 0.6222767 | 0.5859567 | 3 |
| Wong_and_Hofmann_2020 | 0.0618902 | 0.0012199 | 28.6378246 | 0.0750000 | 0.0487805 | 2 |
| Wu_et_al_2013 | 0.1433880 | 0.0004693 | 46.2016834 | 0.2426450 | 0.0441310 | 2 |
| Xu_et_al_2015 | 0.1631094 | 0.0024202 | 20.3312628 | 0.2073162 | 0.1189025 | 2 |
| Xue_and_Ma_2020 | 0.1237757 | 0.0031886 | 19.1650647 | 0.3240480 | -0.1172380 | 12 |
| Yoon_et_al_2019 | -0.1225325 | 0.0448945 | 6.1378861 | 0.1492600 | -0.5357700 | 4 |
| Zhang_and_Kieffer_2014 | 0.5200000 | 0.0078690 | 12.0230393 | 0.7800000 | 0.2600000 | 2 |
| Zhou_et_al_2019 | 0.4148405 | 0.0023819 | 22.7363794 | 0.4392430 | 0.3904380 | 2 |
| Zhu_et_al_2019 | 0.0790000 | 0.0001253 | 92.1683669 | 0.1480000 | 0.0340000 | 4 |
| Ziegeweid_et_al_2008 | 0.2934783 | 0.0009170 | 45.0360055 | 0.3043478 | 0.2826087 | 2 |
d %>%
group_by(ref) %>%
summarise(mean_dARR = mean(dARR),
var_dARR = mean(Var_dARR),
prec = mean(precision),
max_dARR = max(dARR),
min_dARR = min(dARR),
n = n()) %>%
ggplot() +
stat_dots(aes(x=mean_dARR), col="darkcyan") # Plot the results
ggplot(d, aes(dARR)) +
geom_histogram(fill = "darkcyan", col = "black", binwidth = 0.02) +
theme_classic() # Plot the distribution of dARR
ggplot(d) +
stat_slab(aes(x=dARR, fill_ramp=stat(cut_cdf_qi(cdf, .width=c(.5, .8, .95), labels=scales::percent_format()))), side="bottom", scale=0.5, show.legend=F, col="darkcyan") +
stat_dots(aes(x=dARR), alpha=0.8, quantiles=1000, dotsize=1.5, shape=16, show.legend=F, col="darkcyan") # More stylised version of the distribution
Missing data patterns
The package DataExplorer has some useful resources for data exploration
plot_missing(d) # Missing data patterns
Frequency of the different variables
plot_bar(d) # Distribution of categorical variables
plot_histogram(d) # Distribution of continuous variables
Map effect sizes against all extracted moderators
Habitat type
# Plot habitat variation in dARR, with data points scaled by precision
ggplot(data=d, aes(y=dARR, x=habitat, col=habitat,size=precision))+
geom_hline(yintercept = 0,linetype = 2, colour = "black",alpha=0.5)+
geom_jitter(alpha=0.5)+ # display data points with a jitter, so they don't overlap
scale_size_continuous(range = c(1, 7))+ # change scaling of data points
labs(x="Experimental design",y="dARR") + # rename axis labels
theme_classic() # Classic theme
ggbetweenstats(data=d, y=dARR, x=habitat) # The ggbetweenstats function provides a neat visualisation of the data along with descriptive statistics
# Plot habitat variation in precision
ggplot(data=d, aes(y=precision, x=habitat, col=habitat))+
geom_hline(yintercept = 0,linetype = 2, colour = "black",alpha=0.5)+
geom_jitter(alpha=0.5)+
scale_size_continuous(range = c(1, 7))+
labs(x="Experimental design",y="Precision (1/SE)") +
theme_classic() 
ggbetweenstats(data=d, y=precision, x=habitat) 
Taxonomic group
# Plot taxonomic variation in dARR, with data points scaled by precision
ggplot(data = d, aes(y = dARR, x = taxonomic_group, col = taxonomic_group, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "dARR") +
theme_classic()
ggbetweenstats(data = d, y = dARR, x = taxonomic_group)
# Plot taxonomic variation in precision, with data points scaled by precision
ggplot(data = d, aes(y = precision, x = taxonomic_group, col = taxonomic_group)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "Precision (1/SE)") +
theme_classic()
ggbetweenstats(data = d, y = precision, x = taxonomic_group)
Time in a common garden condition
# Plot variation in dARR with re-acclimation time, with data points scaled by
# precision
ggplot(data = d, aes(y = dARR, x = time_common_temp, size = precision), alpha = 0.5) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "age_tested", y = "dARR") +
theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = time_common_temp, y = dARR)
# Plot variation in precision with re-acclimation time.
ggplot(data = d, aes(y = precision, x = time_common_temp), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "age_tested", y = "Precision (1/SE)") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = time_common_temp, y = precision)
Experimental design
# Plot variation in dARR between experimental designs (life history variation
# and persistence)
p_design <- ggplot(data = d, aes(y = dARR, x = exp_design, col = exp_design, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "dARR") +
theme_classic()
# Plot variation in precision between experimental designs (life history
# variation and persistence)
p_design_var <- ggplot(data = d, aes(y = precision, x = exp_design, col = exp_design)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "Precision (1/SE)") +
theme_classic()
# Plot variation in dARR between experimental designs (life history variation
# and persistence), with points coloured for each taxonomic group
p_design_taxa <- ggplot(data = d, aes(y = dARR, x = exp_design, col = taxonomic_group,
size = precision)) + geom_hline(yintercept = 0, linetype = 2, colour = "black",
alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Experimental design", y = "dARR") + theme_classic()
# Plot variation in dARR between experimental designs (life history variation
# and persistence), with points coloured for each metric
p_design_metric <- ggplot(data = d, aes(y = dARR, x = exp_design, col = metric, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "dARR") +
theme_classic()
(p_design + p_design_var)/(p_design_taxa + p_design_metric) # Combine plots
ggbetweenstats(data = d, y = dARR, x = exp_design) # dARR variation between designs
ggbetweenstats(data = d, y = precision, x = exp_design) # Precision variation between designs
Direct vs. persistent effects
# Plot variation in dARR between direct and persistent effect sizes
ggplot(data = d, aes(y = dARR, x = brought_common_temp, col = brought_common_temp,
size = precision)) + geom_hline(yintercept = 0, linetype = 2, colour = "black",
alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Experimental design", y = "dARR") + theme_classic()
ggbetweenstats(data = d, y = dARR, x = brought_common_temp)
# Plot variation in precision between direct and persistent effect sizes
ggplot(data = d, aes(y = precision, x = brought_common_temp, col = brought_common_temp)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "Precision (1/SE)") +
theme_classic()
ggbetweenstats(data = d, y = precision, x = brought_common_temp)
Metric type
# Plot variation in dARR between metric types
p_metric <- ggplot(data = d, aes(y = dARR, x = metric, col = metric, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Metric type", y = "dARR") +
theme_classic()
# Plot variation in precision between metric types
p_metric_prec <- ggplot(data = d, aes(y = precision, x = metric, col = metric)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Metric type", y = "Precision (1/SE)") +
theme_classic()
# Plot variation in dARR between metric types, with different colours for each
# taxomic groups
p_metric_taxa <- ggplot(data = d, aes(y = dARR, x = metric, col = taxonomic_group,
size = precision)) + geom_hline(yintercept = 0, linetype = 2, colour = "black",
alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Metric type", y = "dARR") + theme_classic()
# Plot variation in precision between metric types, with different colours for
# each taxomic groups
p_metric_design <- ggplot(data = d, aes(y = dARR, x = metric, col = exp_design, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Metric type", y = "dARR") +
theme_classic()
(p_metric + p_metric_prec)/(p_metric_taxa + p_metric_design) # Combine plots
ggbetweenstats(data = d, y = dARR, x = metric) # dARR variation between metrics
ggbetweenstats(data = d, y = precision, x = metric) # precision variation between metrics
Heating rate
# Plot variation in dARR with heating rate
ggplot(data = d, aes(y = dARR, x = ramping, size = precision), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "age_tested", y = "dARR") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = ramping, y = dARR)
# Plot variation in precision with heating rate
ggplot(data = d, aes(y = precision, x = ramping), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "age_tested", y = "Precision (1/SE)") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = ramping, y = precision)
Viviparous or Oviparous exposure
Note that external fertilizers were considered as “oviparous” because the temperature treatment was experienced through the egg, instead of through the mothers’ body.
# Plot variation in dARR between viviparous and oviparous animals
ggplot(data = d, aes(y = dARR, x = reproduction_mode, col = reproduction_mode, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "dARR") +
theme_classic()
ggbetweenstats(data = d, y = dARR, x = reproduction_mode)
# Plot variation in precision between viviparous and oviparous animals
ggplot(data = d, aes(y = precision, x = reproduction_mode, col = reproduction_mode)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "Precision (1/SE)") +
theme_classic()
ggbetweenstats(data = d, y = precision, x = reproduction_mode)
Sex
# Plot variation in dARR between sexes
ggplot(data = d, aes(y = dARR, x = sex, col = sex, size = precision)) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Experimental design", y = "dARR") + theme_classic()
ggbetweenstats(data = d, y = dARR, x = sex)
# Plot variation in precision between sexes
ggplot(data = d, aes(y = precision, x = sex, col = sex)) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Experimental design", y = "Precision (1/SE)") + theme_classic()
ggbetweenstats(data = d, y = precision, x = sex)
Age tested
# Plot variation in dARR with the age of testing
ggplot(data = d, aes(y = dARR, x = age_tested, size = precision), alpha = 0.5) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "age_tested", y = "dARR") +
theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = age_tested, y = dARR)
# Plot variation in precision with the age of testing
ggplot(data = d, aes(y = precision, x = age_tested), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "age_tested", y = "Precision (1/SE)") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = precision, y = dARR)
Heat tolerance endpoint
# Plot variation in dARR between heat tolerance endpoints
ggplot(data = d, aes(y = dARR, x = endpoint, col = endpoint, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Experimental design", y = "dARR") +
theme_classic()
ggbetweenstats(data = d, y = dARR, x = endpoint)
# Plot variation in precision between heat tolerance endpoints
ggplot(data = d, aes(y = precision, x = endpoint, col = endpoint)) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Experimental design", y = "Precision (1/SE)") + theme_classic()
ggbetweenstats(data = d, y = precision, x = endpoint)
Hatching origin
# Plot variation in dARR between locations where animals hatched
ggplot(data = d, aes(y = dARR, x = origin_hatching, col = origin_hatching, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Origin", y = "dARR") + theme_classic()
ggbetweenstats(data = d, y = dARR, x = origin_hatching)
# Plot variation in precision between locations where animals hatched
ggplot(data = d, aes(y = precision, x = origin_hatching, col = origin_hatching)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Origin", y = "Precision (1/SE)") +
theme_classic()
ggbetweenstats(data = d, y = precision, x = origin_hatching)
Body mass
# Plot variation in dARR with body mass
ggplot(data = d, aes(y = dARR, x = body_mass, size = precision), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Body mass", y = "dARR") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = body_mass, y = dARR)
# Plot variation in precision with body mass
ggplot(data = d, aes(y = precision, x = body_mass), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Body mass", y = "Precision (1/SE)") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = body_mass, y = precision)
Whether the acclimation temperatures used are independent from housing temperatures prior to acclimation
# Plot variation in dARR depending on whether all temperature acclimation
# treatments were independent from housing temperatures prior to acclimation
ggplot(data = d, aes(y = dARR, x = incubation_independent, col = incubation_independent,
size = precision)) + geom_hline(yintercept = 0, linetype = 2, colour = "black",
alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Origin", y = "dARR") + theme_classic()
ggbetweenstats(data = d, y = dARR, x = incubation_independent)
# Plot variation in precision depending on whether all temperature acclimation
# treatments were independent from housing temperatures prior to acclimation
ggplot(data = d, aes(y = precision, x = incubation_independent, col = incubation_independent)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Origin", y = "Precision (1/SE)") +
theme_classic()
ggbetweenstats(data = d, y = precision, x = incubation_independent)
The time after which survival was assessed in static assays
# Plot variation in dARR depending on experiment duration in static assays
ggplot(data = d, aes(y = dARR, x = set_time, size = precision), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Body mass", y = "dARR") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = set_time, y = dARR)
# Plot variation in precision depending on experiment duration in static assays
ggplot(data = d, aes(y = precision, x = set_time), alpha = 0.5) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Body mass", y = "Precision (1/SE)") + theme_classic() + geom_smooth(method = "lm")
ggscatterstats(data = d, x = set_time, y = precision)
Concerns with data quality
d <- d %>%
mutate(is_concern = ifelse(is.na(major_concerns), "no", "yes")) # Create a new column to indicate whether there are concerns with the data or not
# Plot variation in dARR depending on whether the study has a risk of bias
ggplot(data = d, aes(y = dARR, x = is_concern, col = is_concern, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "Major data concerns", y = "dARR") +
theme_classic()
ggbetweenstats(data = d, x = is_concern, y = dARR)
# Plot variation in precision depending on whether the study has a risk of bias
ggplot(data = d, aes(y = precision, x = is_concern, col = is_concern)) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "Major data concerns", y = "Precision (1/SE)") + theme_classic()
ggbetweenstats(data = d, x = is_concern, y = precision)
Imputed vs original data
# Plot variation in dARR depending on whether the sampling variance was imputed
ggplot(data = d, aes(y = dARR, x = imputed, col = imputed, size = precision)) + geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) + scale_size_continuous(range = c(1,
7)) + labs(x = "SD imputation", y = "dARR") + theme_classic()
ggbetweenstats(data = d, y = dARR, x = imputed)
# Plot variation in precision depending on whether the sampling variance was
# imputed
ggplot(data = d, aes(y = precision, x = imputed, col = Var_dARR, size = precision)) +
geom_hline(yintercept = 0, linetype = 2, colour = "black", alpha = 0.5) + geom_jitter(alpha = 0.5) +
scale_size_continuous(range = c(1, 7)) + labs(x = "SD imputation", y = "Precision (1/SE)") +
theme_classic()
ggbetweenstats(data = d, y = precision, x = imputed)
Phylogeny
Potamilus alatus was not classified in the Open Tree of Life taxonomy. Quick searches in the Integrated Taxonomic Information System for all bivalve species included in our study indicate that Potamilus alatus belongs to the Lampsilini tribe. Species of the genus Villosa, Ligumia, Lampsilis and Ellipsaria all belong to this tribe.
Therefore, this species was arbitrarily added next to Villosa delumbis and polytomies were resolved at random
d <- as.data.frame(d)
d$genus_species = as.factor(d$genus_species) # Convert species name as factor
taxa <- tnrs_match_names(names = levels(d$genus_species), context = "Animals") # Match species name to the Open Tree of Life taxonomy
kable(taxa) # Some species are synonyms, but they don't have >1 match so it's all good.| search_string | unique_name | approximate_match | ott_id | is_synonym | flags | number_matches |
|---|---|---|---|---|---|---|
| acanthochromis polyacanthus | Acanthochromis polyacanthus | FALSE | 100410 | FALSE | incertae_sedis_inherited | 1 |
| acartia tonsa | Acartia tonsa | FALSE | 265973 | FALSE | sibling_higher | 1 |
| acipenser brevirostrum | Acipenser brevirostrum | FALSE | 82746 | FALSE | 1 | |
| acipenser fulvescens | Acipenser fulvescens | FALSE | 698270 | FALSE | 1 | |
| acipenser transmontanus | Acipenser transmontanus | FALSE | 378925 | FALSE | 1 | |
| aedes aegypti | Aedes aegypti | FALSE | 269666 | FALSE | 1 | |
| agosia chrysogaster | Agosia chrysogaster | FALSE | 308519 | FALSE | 1 | |
| alasmidonta varicosa | Alasmidonta varicosa | FALSE | 381439 | FALSE | sibling_higher | 1 |
| amalosia lesueurii | Amalosia lesueurii | FALSE | 295655 | FALSE | 2 | |
| amblema plicata | Amblema plicata | FALSE | 88263 | FALSE | sibling_higher | 1 |
| ambystoma mexicanum | Ambystoma mexicanum | FALSE | 984726 | FALSE | 1 | |
| amphiprion melanopus | Amphiprion melanopus | FALSE | 45635 | FALSE | incertae_sedis_inherited | 1 |
| anisotremus scapularis | Anisotremus scapularis | FALSE | 115727 | FALSE | 1 | |
| anolis sagrei | Anolis sagrei | FALSE | 970153 | FALSE | 1 | |
| anopheles albimanus | Anopheles albimanus | FALSE | 793189 | FALSE | 1 | |
| apostichopus japonicus | Apostichopus japonicus | FALSE | 721060 | FALSE | 1 | |
| argyrosomus regius | Argyrosomus regius | FALSE | 204035 | FALSE | 1 | |
| belonesox belizanus | Belonesox belizanus | FALSE | 913297 | FALSE | 1 | |
| bidyanus bidyanus | Bidyanus bidyanus | FALSE | 600523 | FALSE | 1 | |
| bufo gargarizans | Bufo gargarizans | FALSE | 279549 | FALSE | 1 | |
| busseola fusca | Busseola fusca | FALSE | 535636 | FALSE | 1 | |
| catla catla | Labeo catla | FALSE | 719907 | TRUE | sibling_higher | 1 |
| chilo partellus | Chilo partellus | FALSE | 1062112 | FALSE | 1 | |
| chiloscyllium plagiosum | Chiloscyllium plagiosum | FALSE | 962561 | FALSE | 1 | |
| cirrhinus mrigala | Cirrhinus mrigala | FALSE | 10821 | FALSE | 2 | |
| clupea harengus | Clupea harengus | FALSE | 1005932 | FALSE | 1 | |
| culex pipiens | Culex pipiens | FALSE | 218684 | FALSE | 1 | |
| cydia pomonella | Cydia pomonella | FALSE | 819146 | FALSE | 1 | |
| cyprinodon nevadensis | Cyprinodon nevadensis | FALSE | 617460 | FALSE | 1 | |
| cyprinus carpio | Cyprinus carpio | FALSE | 429083 | FALSE | 1 | |
| danio rerio | Danio rerio | FALSE | 1005914 | FALSE | 1 | |
| daphnia galeata | Daphnia galeata | FALSE | 46786 | FALSE | sibling_higher | 1 |
| daphnia magna | Daphnia magna | FALSE | 668392 | FALSE | sibling_higher | 1 |
| daphnia pulicaria | Daphnia pulicaria | FALSE | 668394 | FALSE | sibling_higher | 1 |
| dicentrarchus labrax | Dicentrarchus labrax | FALSE | 3549 | FALSE | 1 | |
| drosophila melanogaster | Drosophila melanogaster | FALSE | 505714 | FALSE | 1 | |
| drosophila suzukii | Drosophila suzukii | FALSE | 227752 | FALSE | 1 | |
| ellipsaria lineolata | Ellipsaria lineolata | FALSE | 190876 | FALSE | 1 | |
| engystomops pustulosus | Engystomops pustulosus | FALSE | 262675 | FALSE | 1 | |
| enteroctopus megalocyathus | Enteroctopus megalocyathus | FALSE | 746557 | FALSE | 1 | |
| evechinus chloroticus | Evechinus chloroticus | FALSE | 924047 | FALSE | 1 | |
| gambusia affinis | Gambusia affinis | FALSE | 617445 | FALSE | 1 | |
| glossina pallidipes | Glossina pallidipes | FALSE | 753729 | FALSE | 1 | |
| haliotis diversicolor | Haliotis diversicolor | FALSE | 780542 | FALSE | 1 | |
| harmonia axyridis | Harmonia axyridis | FALSE | 258281 | FALSE | 1 | |
| heterodontus portusjacksoni | Heterodontus portusjacksoni | FALSE | 1019118 | FALSE | 1 | |
| hippocampus erectus | Hippocampus erectus | FALSE | 630161 | FALSE | 1 | |
| hodotermes mossambicus | Hodotermes mossambicus | FALSE | 973344 | FALSE | 1 | |
| hypomesus transpacificus | Hypomesus transpacificus | FALSE | 348839 | FALSE | 1 | |
| hypophthalmichthys molitrix | Hypophthalmichthys molitrix | FALSE | 35786 | FALSE | 1 | |
| hypsiboas faber | Boana faber | FALSE | 500272 | TRUE | 1 | |
| ictalurus punctatus | Ictalurus punctatus | FALSE | 701523 | FALSE | 1 | |
| ischnura elegans | Ischnura elegans | FALSE | 253311 | FALSE | 1 | |
| kryptolebias marmoratus | Kryptolebias marmoratus | FALSE | 229709 | FALSE | 1 | |
| labeo rohita | Labeo rohita | FALSE | 160616 | FALSE | sibling_higher | 1 |
| lampropholis coggeri | Lampropholis coggeri | FALSE | 1073190 | FALSE | 1 | |
| lampsilis abrupta | Lampsilis abrupta | FALSE | 1080021 | FALSE | sibling_higher | 1 |
| lampsilis cariosa | Lampsilis cariosa | FALSE | 788597 | FALSE | sibling_higher | 1 |
| lampsilis radiata | Lampsilis radiata | FALSE | 281722 | FALSE | sibling_higher | 1 |
| lampsilis siliquoidea | Lampsilis siliquoidea | FALSE | 765106 | FALSE | sibling_higher | 1 |
| lates calcarifer | Lates calcarifer | FALSE | 437601 | FALSE | incertae_sedis_inherited | 1 |
| lepdomeda copei | Lepidomeda copei | TRUE | 1073687 | FALSE | 1 | |
| leucoraja erinacea | Leucoraja erinacea | FALSE | 493418 | FALSE | 1 | |
| ligumia recta | Ligumia recta | FALSE | 2851633 | FALSE | sibling_higher | 1 |
| limnodynastes peronii | Limnodynastes peronii | FALSE | 276279 | FALSE | sibling_higher | 1 |
| litopenaeus stylirostris | Penaeus stylirostris | FALSE | 957079 | TRUE | 1 | |
| litopenaeus vannamei | Penaeus vannamei | FALSE | 169168 | TRUE | 1 | |
| loligo vulgaris | Loligo vulgaris | FALSE | 1054848 | FALSE | 1 | |
| loxechinus albus | Loxechinus albus | FALSE | 555689 | FALSE | 1 | |
| lutjanus guttatus | Lutjanus guttatus | FALSE | 445054 | FALSE | 1 | |
| macrobrachium rosenbergii | Macrobrachium rosenbergii | FALSE | 708755 | FALSE | 1 | |
| manduca sexta | Manduca sexta | FALSE | 116647 | FALSE | 1 | |
| mauremys mutica | Mauremys mutica | FALSE | 679467 | FALSE | 1 | |
| mauremys reevesii | Mauremys reevesii | FALSE | 829098 | FALSE | 1 | |
| megalonaias nervosa | Megalonaias nervosa | FALSE | 609673 | FALSE | 1 | |
| mesocentrotus franciscanus | Mesocentrotus franciscanus | FALSE | 564712 | FALSE | 1 | |
| metapenaeus affinis | Metapenaeus affinis | FALSE | 675009 | FALSE | 1 | |
| micropterus salmoides | Micropterus salmoides | FALSE | 230493 | FALSE | 1 | |
| morone saxatilis | Morone saxatilis | FALSE | 223669 | FALSE | 1 | |
| moxostoma robustum | Moxostoma robustum | FALSE | 300757 | FALSE | 1 | |
| myxocyprinus asiaticus | Myxocyprinus asiaticus | FALSE | 141409 | FALSE | 1 | |
| myzus persicae | Myzus persicae | FALSE | 785522 | FALSE | 2 | |
| nilaparvata lugens | Nilaparvata lugens | FALSE | 876508 | FALSE | 1 | |
| octopus maya | Octopus maya | FALSE | 237487 | FALSE | 1 | |
| oncorhynchus clarkii | Oncorhynchus clarkii | FALSE | 800432 | FALSE | 1 | |
| oncorhynchus kisutch | Oncorhynchus kisutch | FALSE | 739917 | FALSE | 1 | |
| oncorhynchus mykiss | Oncorhynchus mykiss (species in domain Eukaryota) | FALSE | 165368 | FALSE | 1 | |
| oncorhynchus nerka | Oncorhynchus nerka | FALSE | 165375 | FALSE | 1 | |
| oncorhynchus tshawytscha | Oncorhynchus tshawytscha | FALSE | 730762 | FALSE | 1 | |
| oreochromis niloticus | Oreochromis niloticus | FALSE | 288063 | FALSE | sibling_higher | 1 |
| oryzias melastigma | Oryzias melastigma | FALSE | 179491 | FALSE | 2 | |
| pangasius pangasius | Pangasius pangasius | FALSE | 216174 | FALSE | 1 | |
| parabramis pekinensis | Parabramis pekinensis | FALSE | 837024 | FALSE | 1 | |
| paralichthys californicus | Paralichthys californicus | FALSE | 301358 | FALSE | 1 | |
| pelodiscus sinensis | Pelodiscus sinensis | FALSE | 143391 | FALSE | 1 | |
| penaeus monodon | Penaeus monodon | FALSE | 212713 | FALSE | 1 | |
| petromyzon marinus | Petromyzon marinus | FALSE | 227494 | FALSE | 1 | |
| physalaemus cuvieri | Physalaemus cuvieri | FALSE | 642784 | FALSE | 1 | |
| pimephales promelas | Pimephales promelas | FALSE | 564459 | FALSE | 1 | |
| pleurodema diplolistris | Pleurodema diplolister | FALSE | 806505 | TRUE | sibling_higher | 1 |
| portunus pelagicus | Portunus pelagicus | FALSE | 599991 | FALSE | 1 | |
| potamilus alatus | Potamilus alatus | FALSE | 732215 | FALSE | unplaced_inherited | 1 |
| premnas biaculeatus | Premnas biaculeatus | FALSE | 519096 | FALSE | incertae_sedis_inherited | 1 |
| procypris rabaudi | Procypris rabaudi | FALSE | 328803 | FALSE | 1 | |
| propylea japonica | Propylea japonica | FALSE | 790672 | FALSE | 1 | |
| pseudacris hypochondriaca | Pseudacris hypochondriaca | FALSE | 3619952 | FALSE | 1 | |
| pseudococcus jackbeardsleyi | Pseudococcus jackbeardsleyi | FALSE | 135880 | FALSE | 1 | |
| pseudocrenilabrus multicolor | Pseudocrenilabrus multicolor | FALSE | 338070 | FALSE | 1 | |
| pseudopleuronectes yokohamae | Pseudopleuronectes yokohamae | FALSE | 608831 | FALSE | 1 | |
| pterophyllum scalare | Pterophyllum scalare | FALSE | 220929 | FALSE | 1 | |
| rana temporaria | Rana temporaria | FALSE | 14718 | FALSE | 1 | |
| rhinella granulosa | Rhinella granulosa | FALSE | 962298 | FALSE | 1 | |
| rhinella icterica | Rhinella icterica | FALSE | 745776 | FALSE | 1 | |
| rhinella marina | Rhinella marina | FALSE | 889340 | FALSE | 1 | |
| rhinella ornata | Rhinella ornata | FALSE | 721631 | FALSE | 1 | |
| salmo salar | Salmo salar | FALSE | 688328 | FALSE | 1 | |
| salmo trutta | Salmo trutta | FALSE | 688332 | FALSE | 1 | |
| salvelinus alpinus | Salvelinus alpinus | FALSE | 688324 | FALSE | 1 | |
| salvelinus fontinalis | Salvelinus fontinalis | FALSE | 688339 | FALSE | 1 | |
| salvelinus namaycush | Salvelinus namaycush | FALSE | 114152 | FALSE | 1 | |
| sardina pilchardus | Sardina pilchardus | FALSE | 1027727 | FALSE | 1 | |
| scaphirhynchus albus | Scaphirhynchus albus | FALSE | 962544 | FALSE | 1 | |
| scaphirhynchus platorynchus | Scaphirhynchus platorynchus | FALSE | 904110 | FALSE | 1 | |
| schizopygopsis younghusbandi | Schizopygopsis younghusbandi | FALSE | 203753 | FALSE | 1 | |
| schizothorax kozlovi | Schizothorax kozlovi | FALSE | 709859 | FALSE | 1 | |
| sciaenops ocellatus | Sciaenops ocellatus | FALSE | 433079 | FALSE | 1 | |
| sebastes schlegeli | Sebastes schlegelii | FALSE | 206127 | TRUE | sibling_higher | 1 |
| sesamia calamistis | Sesamia calamistis | FALSE | 112903 | FALSE | 1 | |
| sitobion avenae | Sitobion avenae | FALSE | 1004324 | FALSE | 1 | |
| solea senegalensis | Solea senegalensis | FALSE | 366998 | FALSE | 1 | |
| sparus aurata | Sparus aurata | FALSE | 760723 | FALSE | 1 | |
| sphoeroides annulatus | Sphoeroides annulatus | FALSE | 6406 | FALSE | 1 | |
| thymallus arcticus | Thymallus arcticus | FALSE | 713375 | FALSE | 1 | |
| tigriopus californicus | Tigriopus californicus | FALSE | 461524 | FALSE | 1 | |
| tor putitora | Tor putitora | FALSE | 720179 | FALSE | 1 | |
| trachemys scripta | Trachemys scripta | FALSE | 725741 | FALSE | 1 | |
| villosa delumbis | Villosa delumbis | FALSE | 15452 | FALSE | sibling_higher | 1 |
| xenopus laevis | Xenopus laevis | FALSE | 465096 | FALSE | 1 |
taxa$unique_name <- gsub(" ", "_", taxa$unique_name) # Replace spaces between species name by an underscore
# phylo_tree <- tol_induced_subtree(ott_ids = taxa$ott_id, label_format =
# 'name') # This is a comment because this code cannot run without the
# additional processing below.
# Potamilus alatus (ott732215) is not found in the Open tree taxonomy
d %>%
filter(class == "Bivalvia") %>%
select(order, family, genus_species) %>%
arrange(genus_species) # check all species of bivalves to see where Potamilus alatus sits## order family genus_species
## 1 Unionida Unionidae Alasmidonta varicosa
## 2 Unionida Unionidae Amblema plicata
## 3 Unionida Unionidae Ellipsaria lineolata
## 4 Unionida Unionidae Lampsilis abrupta
## 5 Unionida Unionidae Lampsilis abrupta
## 6 Unionida Unionidae Lampsilis abrupta
## 7 Unionida Unionidae Lampsilis cariosa
## 8 Unionida Unionidae Lampsilis cariosa
## 9 Unionida Unionidae Lampsilis radiata
## 10 Unionida Unionidae Lampsilis siliquoidea
## 11 Unionida Unionidae Lampsilis siliquoidea
## 12 Unionida Unionidae Lampsilis siliquoidea
## 13 Unionida Unionidae Lampsilis siliquoidea
## 14 Unionida Unionidae Ligumia recta
## 15 Unionida Unionidae Ligumia recta
## 16 Unionida Unionidae Megalonaias nervosa
## 17 Unionida Unionidae Potamilus alatus
## 18 Unionida Unionidae Potamilus alatus
## 19 Unionida Unionidae Villosa delumbistaxa.tree <- filter(taxa, ott_id != "732215") # remove Potamilus alatus, as it will be manually added later
phylo_tree <- tol_induced_subtree(ott_ids = taxa.tree$ott_id, label_format = "name") # Generate phylogenetic tree based on the Open Tree of Life taxonomyott_in_tree <- ott_id(taxa.tree)[is_in_tree(ott_id(taxa.tree))] # Make sure that all identifiers are in the taxonomy
phylo_tree <- tol_induced_subtree(ott_ids = ott_in_tree) # Generate tree with all species found in the taxonomyis.binary(phylo_tree) # Check if tree is binary ## [1] FALSEset.seed(1) # Set a seed to resolve politomies at random, and obtain similar results
binary.tree <- multi2di(phylo_tree, random = T) #resolve polytomies at random
binary.tree <- bind.tip(binary.tree, tip.label = "Potamilus_alatus", where = 95) # Add Potamilus alatus to the tree
phylo_branched <- compute.brlen(binary.tree, method = "Grafen", power = 1) # compute branch lengths using Grafen's method
phylo_branched$tip.label <- strip_ott_ids(phylo_branched$tip.label, remove_underscores = FALSE) # remove ott ID from species name to match it to the data set
phylo_matrix <- vcv(phylo_branched, cor = T) # Generate variance covariance matrix to correlate species relatedness
d <- as.data.frame(d)
d <- mutate(d, search_string = decapitalize(genus_species)) # Decapitalise species name to match it to the species name presented in 'taxa'
d <- left_join(d, select(taxa, search_string, unique_name, ott_id), by = "search_string") # Join data sets
d <- d[d$unique_name %in% phylo_branched$tip.label, ] # Check that species names are well matched with the phylogenetic treePlot the tree
Note that the tree presented in Figure 3 of the manuscript was generated using the code in Figures for publication
d.tree <- mutate(d, tip.label = unique_name) # For the correspondence between tree and data
d.tree$tip.label = as.factor(d.tree$tip.label) # Convert to factor
d.tree$tip.label <- gsub("_", " ", d.tree$tip.label) # remove underscores from data
phylo_branched$tip.label <- strip_ott_ids(phylo_branched$tip.label, remove_underscores = TRUE) # remove underscores from tree
summary(d.tree$tip.label) # Need to remove duplicate species for building the tree## Length Class Mode
## 1089 character characterd.tree <- d.tree %>%
group_by(tip.label) %>%
mutate(n_es = sum(n = n())) %>%
ungroup() # Count the number of effect sizes for each species
d.tree <- mutate(d.tree, log_es = log10(n_es)) # Log the number of effect sizes, if needed for visualisations
d.tree_plot <- distinct(d.tree, tip.label, .keep_all = TRUE) # Only keep one row per species
d.tree_plot <- as.data.frame(d.tree_plot) # Convert tibble to data frame
d.tree_plot$phylum <- as.character(d.tree_plot$phylum) # Convert phylum to character
d.tree_plot <- select(d.tree_plot, tip.label, phylum, class, family, exp_design,
metric, n_es, log_es, dARR) # important: select only the columns you need, otherwise this does not run
tree_for_plot <- drop.tip(phylo_branched, setdiff(phylo_branched$tip.label, phylo_branched$tip.label)) # Make sure to keep all matching observations
design_dat <- d.tree %>%
group_by(tip.label) %>%
summarise(direct = brought_common_temp == "no", persistent = brought_common_temp ==
"yes", CTmax = metric == "CTmax", LT50 = metric == "LT50") # Summarise, for each species, whether they tested direct effects, persistent effects, CTmax, or LT50
design_dat <- distinct(design_dat) # Only keep unique rows
design_dat$direct = as.numeric(design_dat$direct) # convert TRUE/FALSE to binary values
design_dat$persistent = as.numeric(design_dat$persistent) # convert TRUE/FALSE to binary values
design_dat$CTmax = as.numeric(design_dat$CTmax) #convert TRUE/FALSE to binary values
design_dat$LT50 = as.numeric(design_dat$LT50) #convert TRUE/FALSE to binary values
design_dat <- design_dat %>%
group_by(tip.label) %>%
summarise(direct = sum(direct), persistent = sum(persistent), CTmax = sum(CTmax),
LT50 = sum(LT50)) # calculate the sum for each species (i.e. if 1, the species has the given design or metric)
# In some cases, values were repeated, so need to replace '2' values by '1'
design_dat$direct[design_dat$direct == "2"] <- "1"
design_dat$persistent[design_dat$persistent == "2"] <- "1"
design_dat$CTmax[design_dat$CTmax == "2"] <- "1"
design_dat$direct = as.factor(design_dat$direct) # convert back to factor for the plot
design_dat$persistent = as.factor(design_dat$persistent)
design_dat$CTmax = as.factor(design_dat$CTmax)
design_dat$LT50 = as.factor(design_dat$LT50)
design_dat <- mutate(design_dat, persistence = ifelse(direct == "1" & persistent ==
"1", "both", ifelse(direct == "1" & persistent == "0", "direct", "persistent"))) # If both direct and persistent, indicate 'both', if only direct, indicate 'direct', otherwise indicate 'persistent'
design_dat <- mutate(design_dat, metrics = ifelse(CTmax == "1" & LT50 == "1", "both",
ifelse(CTmax == "1" & LT50 == "0", "CTmax", "LT50"))) # If both CTmax and LT50, indicate 'both', if only CTmax, indicate 'CTmax', otherwise indicate 'LT50'
d.tree_plot <- left_join(d.tree_plot, design_dat, by = "tip.label") # Join this information to the rest of the data we want to plot
p <- ggtree(tree_for_plot, layout = "circular", lwd = 0.75) # Circular tree
p <- p %<+% d.tree_plot # link plot to data
p2 <- p + geom_fruit(geom = geom_tile, mapping = aes(fill = metrics), width = 0.07,
offset = 0.085, col = "gray30") + scale_fill_manual(values = c("gray70", "white",
"black")) # Create tiles to indicate which metric was used for this species
p3 <- p2 + new_scale_fill() + geom_fruit(geom = geom_tile, mapping = aes(fill = persistence),
offset = 0.1, width = 0.07, col = "gray30") + scale_fill_manual(values = c("#7570B3",
"#D95F02", "#1B9E77")) # Create tiles to indicate whether direct or persistent effects were assessed
d.tree_plot$n_es = as.numeric(d.tree_plot$n_es) # Convert to numeric
p4 <- p3 + new_scale_fill() + geom_fruit(geom = geom_bar, mapping = aes(x = n_es,
fill = class), stat = "identity", col = "gray1", orientation = "y", axis.params = list(axis = "x",
text.angle = -45, hjust = 0, text.size = 3), border = 1.2, grid.params = list(alpha = 0.35),
offset = 0.085, pwidth = 0.55, alpha = 0.8) + scale_fill_manual(values = c("#00BBDB",
"darkslategray4", "darkorange", "chartreuse", "indianred2", "lightsteelblue2",
"darkorchid", "darkseagreen1", "deeppink", "darkred", "forestgreen", "darkolivegreen2",
"#39568CFF", "darkgoldenrod2")) # Display number of effect sizes
p4
Account for treatments re-used in multiple comparisons
In many cases, authors used multiple temperature treatments that can be compared. In this case, we calculated our effect sizes on stepwise comparisons (e.g., 15C-17C; 17C-19C, 19C-22C). Doing such creates a source of non-independence that needs to be accounted for. In our case, sampling errors from treatments involved in multiple comparisons were correlated (using a conservative r = 0.5) with a variance covariance matrix using the make_VCV_matrix function from the metaAidR package.
names(d) <- make.names(names(d))
d <- as.data.frame(d) # Important: only works with data frames!
VCV_dARR <- make_VCV_matrix(d, V = "Var_dARR", cluster = "shared_trt_ID", obs = "es_ID")Sample sizes
d %>%
summarise(
# Number of studies, species and effect sizes
`Studies` = n_distinct(study_ID),
`Species` = n_distinct(genus_species),
`Effect sizes` = n_distinct(es_ID),
# Number of studies, species, and effect sizes for direct and persistent designs
`Studies (Direct)` = n_distinct(study_ID[brought_common_temp=="no"]),
`Studies (Persistent)` = n_distinct(study_ID[brought_common_temp=="yes"]),
`Species (Direct)` = n_distinct(genus_species[brought_common_temp=="no"]),
`Species (Persistent)` = n_distinct(genus_species[brought_common_temp=="yes"]),
`Effect sizes (Direct)` = n_distinct(es_ID[brought_common_temp=="no"]),
`Effect sizes (Persistent)` = n_distinct(es_ID[brought_common_temp=="yes"]),
# Number of studies, species, and effect sizes for each habitat
`Studies (Aquatic)` = n_distinct(study_ID[habitat=="aquatic"]),
`Studies (Terrestrial)` = n_distinct(study_ID[habitat=="terrestrial"]),
`Species (Aquatic)` = n_distinct(genus_species[habitat=="aquatic"]),
`Species (Terrestrial)` = n_distinct(genus_species[habitat=="terrestrial"]),
`Effect sizes (Aquatic)` = n_distinct(es_ID[habitat=="aquatic"]),
`Effect sizes (Terrestrial)` = n_distinct(es_ID[habitat=="terrestrial"]),
# Number of studies, species, and effect sizes for each taxonomic group
`Studies (Fishes)` = n_distinct(study_ID[taxonomic_group=="fish"]),
`Studies (Reptiles)` = n_distinct(study_ID[taxonomic_group=="reptile"]),
`Studies (Amphibians)` = n_distinct(study_ID[taxonomic_group=="amphibian"]),
`Studies (Aquatic invertebrates)` = n_distinct(study_ID[taxonomic_group=="aquatic invertebrate"]),
`Studies (Terrestrial invertebrates)` = n_distinct(study_ID[taxonomic_group=="terrestrial invertebrate"]),
`Studies (Arthropods)` = n_distinct(study_ID[phylum=="Arthropoda"]),
`Studies (Echinoderms)` = n_distinct(study_ID[phylum=="Echinodermata"]),
`Studies (Molluscs)` = n_distinct(study_ID[phylum=="Mollusca"]),
`Species (Fishes)` = n_distinct(genus_species[taxonomic_group=="fish"]),
`Species (Reptiles)` = n_distinct(genus_species[taxonomic_group=="reptile"]),
`Species (Amphibians)` = n_distinct(genus_species[taxonomic_group=="amphibian"]),
`Species (Aquatic invertebrates)` = n_distinct(genus_species[taxonomic_group=="aquatic invertebrate"]),
`Species (Terrestrial invertebrates)` = n_distinct(genus_species[taxonomic_group=="terrestrial invertebrate"]),
`Species (Arthropods)` = n_distinct(genus_species[phylum=="Arthropoda"]),
`Species (Echinoderms)` = n_distinct(genus_species[phylum=="Echinodermata"]),
`Species (Molluscs)` = n_distinct(genus_species[phylum=="Mollusca"]),
`Effect sizes (Fishes)` = n_distinct(es_ID[taxonomic_group=="fish"]),
`Effect sizes (Reptiles)` = n_distinct(es_ID[taxonomic_group=="reptile"]),
`Effect sizes (Amphibians)` = n_distinct(es_ID[taxonomic_group=="amphibian"]),
`Effect sizes (Aquatic invertebrates)` = n_distinct(es_ID[taxonomic_group=="aquatic invertebrate"]),
`Effect sizes (Terrestrial invertebrates)` = n_distinct(es_ID[taxonomic_group=="terrestrial invertebrate"]),
`Effect sizes (Arthropods)` = n_distinct(es_ID[phylum=="Arthropoda"]),
`Effect sizes (Echinoderms)` = n_distinct(es_ID[phylum=="Echinodermata"]),
`Effect sizes (Molluscs)` = n_distinct(es_ID[phylum=="Mollusca"]),
# Number of studies, species and effect sizes for each experimental design
`Studies (Design A)` = n_distinct(study_ID[exp_design=="A"]),
`Studies (Design B)` = n_distinct(study_ID[exp_design=="B"]),
`Studies (Design C)` = n_distinct(study_ID[exp_design=="C"]),
`Studies (Design D)` = n_distinct(study_ID[exp_design=="D"]),
`Studies (Design E)` = n_distinct(study_ID[exp_design=="E"]),
`Studies (Design F)` = n_distinct(study_ID[exp_design=="F"]),
`Species (Design A)` = n_distinct(genus_species[exp_design=="A"]),
`Species (Design B)` = n_distinct(genus_species[exp_design=="B"]),
`Species (Design C)` = n_distinct(genus_species[exp_design=="C"]),
`Species (Design D)` = n_distinct(genus_species[exp_design=="D"]),
`Species (Design E)` = n_distinct(genus_species[exp_design=="E"]),
`Species (Design F)` = n_distinct(genus_species[exp_design=="F"]),
`Effect sizes (Design A)` = n_distinct(es_ID[exp_design=="A"]),
`Effect sizes (Design B)` = n_distinct(es_ID[exp_design=="B"]),
`Effect sizes (Design C)` = n_distinct(es_ID[exp_design=="C"]),
`Effect sizes (Design D)` = n_distinct(es_ID[exp_design=="D"]),
`Effect sizes (Design E)` = n_distinct(es_ID[exp_design=="E"]),
`Effect sizes (Design F)` = n_distinct(es_ID[exp_design=="F"]),
# Number of studies, species and effect sizes for each metric type
`Studies (CTmax)` = n_distinct(study_ID[metric=="CTmax"]),
`Studies (LT50)` = n_distinct(study_ID[metric=="LT50"]),
`Species (CTmax)` = n_distinct(genus_species[metric=="CTmax"]),
`Species (LT50)` = n_distinct(genus_species[metric=="LT50"]),
`Effect sizes (CTmax)` = n_distinct(es_ID[metric=="CTmax"]),
`Effect sizes (LT50)` = n_distinct(es_ID[metric=="LT50"]),
) -> table_sample_sizes
table_sample_sizes<-t(table_sample_sizes)
colnames(table_sample_sizes)<-"n (sample size)"
kable(table_sample_sizes) %>% kable_styling("striped", position="center") # Stylise table | n (sample size) | |
|---|---|
| Studies | 150 |
| Species | 138 |
| Effect sizes | 1089 |
| Studies (Direct) | 126 |
| Studies (Persistent) | 26 |
| Species (Direct) | 121 |
| Species (Persistent) | 23 |
| Effect sizes (Direct) | 866 |
| Effect sizes (Persistent) | 223 |
| Studies (Aquatic) | 128 |
| Studies (Terrestrial) | 22 |
| Species (Aquatic) | 118 |
| Species (Terrestrial) | 20 |
| Effect sizes (Aquatic) | 929 |
| Effect sizes (Terrestrial) | 160 |
| Studies (Fishes) | 83 |
| Studies (Reptiles) | 9 |
| Studies (Amphibians) | 10 |
| Studies (Aquatic invertebrates) | 29 |
| Studies (Terrestrial invertebrates) | 19 |
| Studies (Arthropods) | 35 |
| Studies (Echinoderms) | 5 |
| Studies (Molluscs) | 8 |
| Species (Fishes) | 68 |
| Species (Reptiles) | 7 |
| Species (Amphibians) | 14 |
| Species (Aquatic invertebrates) | 31 |
| Species (Terrestrial invertebrates) | 18 |
| Species (Arthropods) | 30 |
| Species (Echinoderms) | 4 |
| Species (Molluscs) | 15 |
| Effect sizes (Fishes) | 623 |
| Effect sizes (Reptiles) | 27 |
| Effect sizes (Amphibians) | 71 |
| Effect sizes (Aquatic invertebrates) | 221 |
| Effect sizes (Terrestrial invertebrates) | 147 |
| Effect sizes (Arthropods) | 292 |
| Effect sizes (Echinoderms) | 13 |
| Effect sizes (Molluscs) | 63 |
| Studies (Design A) | 104 |
| Studies (Design B) | 5 |
| Studies (Design C) | 21 |
| Studies (Design D) | 7 |
| Studies (Design E) | 8 |
| Studies (Design F) | 12 |
| Species (Design A) | 109 |
| Species (Design B) | 5 |
| Species (Design C) | 13 |
| Species (Design D) | 6 |
| Species (Design E) | 7 |
| Species (Design F) | 11 |
| Effect sizes (Design A) | 700 |
| Effect sizes (Design B) | 20 |
| Effect sizes (Design C) | 146 |
| Effect sizes (Design D) | 92 |
| Effect sizes (Design E) | 76 |
| Effect sizes (Design F) | 55 |
| Studies (CTmax) | 132 |
| Studies (LT50) | 30 |
| Species (CTmax) | 118 |
| Species (LT50) | 35 |
| Effect sizes (CTmax) | 863 |
| Effect sizes (LT50) | 226 |
kable(d %>% group_by(study_ID) %>% # Calculate the number of effect sizes per study
summarise(n = n_distinct(es_ID)) %>%
ungroup() %>%
summarise(`n effect sizes per study` = mean(n),
`sd effect sizes per study`= sd (n),
`min effect sizes per study` = min(n),
`max effect sizes per study`= max(n))) %>%
kable_styling("striped", position="center") | n effect sizes per study | sd effect sizes per study | min effect sizes per study | max effect sizes per study |
|---|---|---|---|
| 7.26 | 9.627582 | 1 | 80 |
# Risk of bias
d.concern<-mutate(d,
Elliott=ifelse(major_concerns=="Elliot's hybrid methodology. Sample size unclear."|major_concerns=="Elliott's hybrid methodology"|major_concerns=="Elliott's hybrid methodology."|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 25C"|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 10C and 15C."| major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 15C"| major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 15C."|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 25C."|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 27C."|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 5C."|major_concerns=="Elliott's hybrid methodology. SD missing."|major_concerns=="hybrid between LT50 and CTmax", study_ID, NA),
Missing_est_SD=ifelse(major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 25C"|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 10C and 15C."| major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 15C"| major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 15C."|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 25C."|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 27C."|major_concerns=="Elliott's hybrid methodology. SD missing for animals acclimated to 5C."|major_concerns=="Elliott's hybrid methodology. SD missing."|major_concerns=="Missing sd."|major_concerns=="SD missing"|major_concerns=="SD missing."|major_concerns=="SD possibly under-estimated."|major_concerns=="SE was taken instead of the SD reported."|major_concerns=="SE was taken instead of the SD reported."|major_concerns=="Standard deviation calculated as 0 so it was assigned as \"NA\"."|major_concerns=="Standard deviation was reported as 0 so it was assigned as \"NA\""|major_concerns=="Standard error missing."|major_concerns=="Standard error probably under- or over-estimated because the number of animals at each test temperatures was not clearly stated and is used for calculating LT50."|imputed=="yes", study_ID, NA),
Unclear_N=ifelse(major_concerns=="Elliot's hybrid methodology. Sample size unclear."|major_concerns=="N missing", study_ID, NA),
Tolerance_landscape=ifelse(major_concerns=="The CTmax was estimated from death times at different temperatures (thermal tolerance landscapes)."|major_concerns=="The CTmax was estimated from death times at different temperatures (thermal tolerance landscapes).\\", study_ID, NA),
Data_similar=ifelse(major_concerns=="Data very similar to study_ID 69.", study_ID, NA),
High_SE=ifelse(major_concerns=="Very high standard error because low sample size (4 test temperatures) in the 28C acclimated group.", study_ID, NA),
Short_test=ifelse(major_concerns=="Very short set_time at the test temperatures.", study_ID, NA)
)
table_bias<- d.concern %>%
summarise(
`Studies (data similar)` = n_distinct(Data_similar, na.rm=T),
`Studies (Elliott's hybrid)` = n_distinct(Elliott, na.rm=T),
`Studies (Missing or estimated SD)`= n_distinct(Missing_est_SD, na.rm=T),
`Studies (Unclear sample size)`= n_distinct(Unclear_N, na.rm=T),
`Studies (Tolerance landscape)`= n_distinct(Tolerance_landscape, na.rm=T),
`Studies (High standard error)`= n_distinct(High_SE, na.rm=T),
`Studies (Short test time)`= n_distinct(Short_test, na.rm=T)
)
table_bias<-t(table_bias)
colnames(table_bias)<-"n (sample size)"
kable(table_bias) %>% kable_styling("striped", position="center") | n (sample size) | |
|---|---|
| Studies (data similar) | 1 |
| Studies (Elliott’s hybrid) | 4 |
| Studies (Missing or estimated SD) | 16 |
| Studies (Unclear sample size) | 3 |
| Studies (Tolerance landscape) | 4 |
| Studies (High standard error) | 1 |
| Studies (Short test time) | 1 |
Save processed data
d$phylogeny <- d$unique_name # Rename 'unique_name' to 'phylogeny' for the models
save(VCV_dARR, phylo_matrix, file = here("RData", "Var_and_phylo_matrices.RData"))
write_csv(d, file = here("data", "Processed_data.csv"))
load(here("Rdata", "Var_and_phylo_matrices.RData"))
d.raw <- read_csv("data/Processed_data.csv")Choose random effect structure
Study ID and species ID were not included in the same model because their levels overlapped too much. We kept “species ID” because it allows to decompose phylogenetic and non-phylogenetic species effects, as recommended in Cinar et al. 2021
Because population_ID was explaining virtually no variance, they were removed from further models.
Cohort_ID and shared_treatment_ID were not included as random effects because the non-independence arising from shared measurements and treatments reused in multiple comparisons were already accounted for in the calculation of effect sizes (see equation 4 and 5) and associated sampling variances (see calculation of VCV_dARR).
Note that all models are relatively long to run, but they are saved in the folder models and code for loading each model is presented at the end of each section in the .Rmd file.
d<-as.data.frame(d) #
random_effect_mod<- rma.mv(dARR~1, # Intercept only
V=VCV_dARR,
method="REML", # Restricted maximum likelihood
test="t", # Test statistics using a t distribution
dfs="contain", # Contain degrees of freedom (recommended)
random=list(~1|species_ID,
~1|population_ID,
~1|phylogeny,
~1|es_ID),
R = list(phylogeny = phylo_matrix), # Assign the phylogenetic relatedness matrix to "phylogeny"
data=d)
summary(random_effect_mod) # Summary of the model##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -226.3154 452.6309 462.6309 487.5913 462.6863
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0095 0.0974 138 no species_ID no
## sigma^2.2 0.0000 0.0000 239 no population_ID no
## sigma^2.3 0.0249 0.1578 138 no phylogeny yes
## sigma^2.4 0.0605 0.2460 1089 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 1088) = 63442.0311, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.1898 0.0882 2.1525 137 0.0331 0.0154 0.3641 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Custom functions
Run models
run.model<-function(data,formula){
data<-as.data.frame(data) # convert data set into a data frame to calculate VCV matrix
VCV<-make_VCV_matrix(data, V="Var_dARR", cluster="shared_trt_ID", obs="es_ID") # create VCV matrix for the specified data
mod<- rma.mv(yi=dARR,
V=VCV, # run the model, as described earlier
mods= formula,
method="REML",
test="t",
dfs="contain",
random=list(~1|species_ID,
~1|phylogeny,
~1|es_ID),
R = list(phylogeny = phylo_matrix),
data=data,
sparse=TRUE) # To reduce model run time
return(mod)
}Plot orchaRd plots
my.orchard<- function (object, mod = "1", group, data, xlab, N = "none",
alpha = 0.5, angle = 90, cb = FALSE, k = TRUE, g = TRUE,
trunk.size = 7, branch.size = 2, twig.size = 0.8, whisker, transfm = c("none", # increased point size, branch size, and added a whisker argument
"tanh"), condition.lab = "Condition", legend.pos = c("bottom.right",
"bottom.left", "top.right", "top.left",
"top.out", "bottom.out"), k.pos = c("right",
"left"))
{
transfm <- match.arg(transfm)
if (any(class(object) %in% c("rma.mv", "rma"))) {
if (mod != "1") {
results <- orchaRd::mod_results(object, mod, group,
data)
}
else {
results <- orchaRd::mod_results(object, mod = "1",
group, data)
}
}
if (any(class(object) %in% c("orchard"))) {
results <- object
}
mod_table <- results$mod_table
data <- results$data
data$moderator <- factor(data$moderator, levels = mod_table$name,
labels = mod_table$name)
data$scale <- (1/sqrt(data[, "vi"]))
legend <- "Precision (1/SE)"
if (any(N != "none")) {
data$scale <- N
legend <- "Sample Size (N)"
}
if (transfm == "tanh") {
cols <- sapply(mod_table, is.numeric)
mod_table[, cols] <- Zr_to_r(mod_table[, cols])
data$yi <- Zr_to_r(data$yi)
label <- xlab
}
else {
label <- xlab
}
mod_table$K <- as.vector(by(data, data[, "moderator"],
function(x) length(x[, "yi"])))
mod_table$g <- as.vector(num_studies(data, moderator, stdy)[,
2])
group_no <- length(unique(mod_table[, "name"]))
cbpl <- c("#88CCEE", "#CC6677", "#DDCC77",
"#117733", "#332288", "#AA4499", "#44AA99",
"#999933", "#882255", "#661100", "#6699CC",
"#888888", "#E69F00", "#56B4E9", "#009E73",
"#F0E442", "#0072B2", "#D55E00", "#CC79A7",
"#999999")
if (names(mod_table)[2] == "condition") {
condition_no <- length(unique(mod_table[, "condition"]))
plot <- ggplot2::ggplot() + ggbeeswarm::geom_quasirandom(data = data,
ggplot2::aes(y = yi, x = moderator, size = scale,
colour = moderator), alpha = alpha) + ggplot2::geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = alpha) +
ggplot2::geom_linerange(data = mod_table, ggplot2::aes(x = name,
ymin = lowerCL, ymax = upperCL), size = branch.size,
position = ggplot2::position_dodge2(width = 0.3)) +
ggplot2::geom_pointrange(data = mod_table, ggplot2::aes(y = estimate,
x = name, ymin = lowerPR, ymax = upperPR, shape = as.factor(condition),
fill = name), size = twig.size, position = ggplot2::position_dodge2(width = 0.3),
fatten = trunk.size) + ggplot2::scale_shape_manual(values = 20 +
(1:condition_no)) + ggplot2::coord_flip() + ggplot2::theme_bw() +
ggplot2::guides(fill = "none", colour = "none") +
ggplot2::theme(legend.position = c(0, 1), legend.justification = c(0,
1)) + ggplot2::theme(legend.title = ggplot2::element_text(size = 9)) +
ggplot2::theme(legend.direction = "horizontal") +
ggplot2::theme(legend.background = ggplot2::element_blank()) +
ggplot2::labs(y = label, x = "", size = legend) +
ggplot2::labs(shape = condition.lab) + ggplot2::theme(axis.text.y = ggplot2::element_text(size = 10,
colour = "black", hjust = 0.5, angle = angle))
}
else {
plot <- ggplot2::ggplot() + ggbeeswarm::geom_quasirandom(data = data,
ggplot2::aes(y = yi, x = moderator, size = scale,
fill = moderator), alpha = alpha, width=0.4, pch=21, stroke=1.1, col="black") + # Change point shape (21, with black borders)
ggplot2::geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.3, lwd=1.3) + # Change thickness 0 line
ggplot2::geom_errorbar(data = mod_table, ggplot2::aes(x = name,
ymin = lowerCL, ymax = upperCL), size = branch.size, width= whisker) + # Added variable whisker size
ggplot2::geom_pointrange(data = mod_table, ggplot2::aes(y = estimate,
x = name, ymin = lowerPR, ymax = upperPR, fill = name),
size = twig.size, fatten = trunk.size, shape = 23, stroke=2.2) + # Change point shape
scale_size_continuous(range = c(1, 14))+ # change point scaling
ggplot2::coord_flip() +
ggplot2::theme_bw() +
ggplot2::guides(fill = "none", colour = "none") +
ggplot2::theme(text=element_text(size=26, colour="black"))+ # Change font size
ggplot2::theme(legend.title = ggplot2::element_text(size = 16)) + # Increased font legend title
ggplot2::theme(legend.text = ggplot2::element_text(size = 14)) +
ggplot2::theme(legend.direction = "horizontal") +
ggplot2::theme(legend.background = ggplot2::element_blank()) +
ggplot2::labs(y = label, x = "", size = legend) +
ggplot2::theme(axis.text.y = ggplot2::element_text(size = 20,
colour = "black", hjust = 0.5, angle = angle)) + # Increased size title axis label
ggplot2::theme(axis.text.x = ggplot2::element_text(size = 20)) + # Increase size axis ticks
ggplot2::theme(panel.border = element_rect(colour = "black", fill=NA, size=1.3))
}
if (legend.pos == "bottom.right") {
plot <- plot + ggplot2::theme(legend.position = c(1,
0), legend.justification = c(1, 0))
}
else if (legend.pos == "bottom.left") {
plot <- plot + ggplot2::theme(legend.position = c(0,
0), legend.justification = c(0, 0))
}
else if (legend.pos == "top.right") {
plot <- plot + ggplot2::theme(legend.position = c(1,
1), legend.justification = c(1, 1))
}
else if (legend.pos == "top.left") {
plot <- plot + ggplot2::theme(legend.position = c(0,
1), legend.justification = c(0, 1))
}
else if (legend.pos == "top.out") {
plot <- plot + ggplot2::theme(legend.position = "top")
}
else if (legend.pos == "bottom.out") {
plot <- plot + ggplot2::theme(legend.position = "bottom")
}
if (cb == TRUE) {
plot <- plot + ggplot2::scale_fill_manual(values = cbpl) +
ggplot2::scale_colour_manual(values = cbpl)
}
if (k == TRUE && g == FALSE && k.pos == "right") {
plot <- plot + ggplot2::annotate("text", y = (max(data$yi) +
(max(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no]), # Size changed to 5.5
parse = TRUE, hjust = "right", size = 6.5)
}
else if (k == TRUE && g == FALSE && k.pos == "left") {
plot <- plot + ggplot2::annotate("text", y = (min(data$yi) +
(min(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no]),
parse = TRUE, hjust = "left", size = 6.5) # Size changed to 5.5
}
else if (k == TRUE && g == TRUE && k.pos == "right") {
plot <- plot + ggplot2::annotate("text", y = (max(data$yi) +
(max(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no],
" (", mod_table$g[1:group_no], ")"),
parse = TRUE, hjust = "right", size = 6.5) # Size changed to 5.5
}
else if (k == TRUE && g == TRUE && k.pos == "left") {
plot <- plot + ggplot2::annotate("text", y = (min(data$yi) +
(min(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no],
" (", mod_table$g[1:group_no], ")"),
parse = TRUE, hjust = "left", size = 6.5) # Size changed to 5.5
}
return(plot)
}Plot continuous moderator
plot_continuous<-function(data, model, moderator, xlab){
pred<-predict.rma(model) # Generate predictions from the model
data<- data %>% mutate(fit=pred$pred, # Mutate predicted values and prediction/confidence intervals to the dataset
ci.lb=pred$ci.lb,
ci.ub=pred$ci.ub,
pr.lb=pred$cr.lb,
pr.ub=pred$cr.ub)
plot<-ggplot(data=data, aes(x = moderator, y = dARR)) + # Plot the results
geom_ribbon(aes(ymin = pr.lb, ymax = pr.ub, color = NULL), alpha = .1) + # Shaded area for prediction intervals
geom_ribbon(aes(ymin = ci.lb, ymax = ci.ub, color = NULL), alpha = .3) + # Shaded area for confidence intervals
geom_point(aes(size=precision), shape=21, alpha=0.7, fill="sienna1", col="gray25",stroke=1) + # Plot scaled data points
geom_line(aes(y = fit), size = 1.5)+ # Display predicted regression line
labs(x = xlab, y = "dARR", size = "Precison (1/SE)") +
theme_bw() + # Black and white theme
scale_size_continuous(range=c(1,12)) + # CHange extent of point scaling
geom_hline(yintercept = 0,linetype = 2, colour = "black",alpha=0.5)+ # horizontal line at dARR = 0
theme(text = element_text(size = 18, colour = "black", hjust = 0.5), # change font sizes and legend position
legend.text=element_text(size=14),
legend.position=c(0,0),
legend.justification = c(0,0),
legend.background = element_blank(),
legend.direction="horizontal",
legend.title = element_text(size=15),
panel.border=element_rect(colour="black", fill=NA, size=1.2))
return(plot)
}my.orchard<- function (object, mod = "1", group, data, xlab, N = "none",
alpha = 0.5, angle = 90, cb = FALSE, k = TRUE, g = TRUE,
trunk.size = 7, branch.size = 2, twig.size = 0.8, whisker, transfm = c("none", # increased point size, branch size, and added a whisker argument
"tanh"), condition.lab = "Condition", legend.pos = c("bottom.right",
"bottom.left", "top.right", "top.left",
"top.out", "bottom.out"), k.pos = c("right",
"left"))
{
transfm <- match.arg(transfm)
if (any(class(object) %in% c("rma.mv", "rma"))) {
if (mod != "1") {
results <- orchaRd::mod_results(object, mod, group,
data)
}
else {
results <- orchaRd::mod_results(object, mod = "1",
group, data)
}
}
if (any(class(object) %in% c("orchard"))) {
results <- object
}
mod_table <- results$mod_table
data <- results$data
data$moderator <- factor(data$moderator, levels = mod_table$name,
labels = mod_table$name)
data$scale <- (1/sqrt(data[, "vi"]))
legend <- "Precision (1/SE)"
if (any(N != "none")) {
data$scale <- N
legend <- "Sample Size (N)"
}
if (transfm == "tanh") {
cols <- sapply(mod_table, is.numeric)
mod_table[, cols] <- Zr_to_r(mod_table[, cols])
data$yi <- Zr_to_r(data$yi)
label <- xlab
}
else {
label <- xlab
}
mod_table$K <- as.vector(by(data, data[, "moderator"],
function(x) length(x[, "yi"])))
mod_table$g <- as.vector(num_studies(data, moderator, stdy)[,
2])
group_no <- length(unique(mod_table[, "name"]))
cbpl <- c("#88CCEE", "#CC6677", "#DDCC77",
"#117733", "#332288", "#AA4499", "#44AA99",
"#999933", "#882255", "#661100", "#6699CC",
"#888888", "#E69F00", "#56B4E9", "#009E73",
"#F0E442", "#0072B2", "#D55E00", "#CC79A7",
"#999999")
if (names(mod_table)[2] == "condition") {
condition_no <- length(unique(mod_table[, "condition"]))
plot <- ggplot2::ggplot() + ggbeeswarm::geom_quasirandom(data = data,
ggplot2::aes(y = yi, x = moderator, size = scale,
colour = moderator), alpha = alpha) + ggplot2::geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = alpha) +
ggplot2::geom_linerange(data = mod_table, ggplot2::aes(x = name,
ymin = lowerCL, ymax = upperCL), size = branch.size,
position = ggplot2::position_dodge2(width = 0.3)) +
ggplot2::geom_pointrange(data = mod_table, ggplot2::aes(y = estimate,
x = name, ymin = lowerPR, ymax = upperPR, shape = as.factor(condition),
fill = name), size = twig.size, position = ggplot2::position_dodge2(width = 0.3),
fatten = trunk.size) + ggplot2::scale_shape_manual(values = 20 +
(1:condition_no)) + ggplot2::coord_flip() + ggplot2::theme_bw() +
ggplot2::guides(fill = "none", colour = "none") +
ggplot2::theme(legend.position = c(0, 1), legend.justification = c(0,
1)) + ggplot2::theme(legend.title = ggplot2::element_text(size = 9)) +
ggplot2::theme(legend.direction = "horizontal") +
ggplot2::theme(legend.background = ggplot2::element_blank()) +
ggplot2::labs(y = label, x = "", size = legend) +
ggplot2::labs(shape = condition.lab) + ggplot2::theme(axis.text.y = ggplot2::element_text(size = 10,
colour = "black", hjust = 0.5, angle = angle))
}
else {
plot <- ggplot2::ggplot() + ggbeeswarm::geom_quasirandom(data = data,
ggplot2::aes(y = yi, x = moderator, size = scale,
fill = moderator), alpha = alpha, width=0.4, pch=21, stroke=1.1, col="black") + # Change point shape (21, with black borders)
ggplot2::geom_hline(yintercept = 0,
linetype = 2, colour = "black", alpha = 0.3, lwd=1.3) + # Change thickness 0 line
ggplot2::geom_errorbar(data = mod_table, ggplot2::aes(x = name,
ymin = lowerCL, ymax = upperCL), size = branch.size, width= whisker) + # Added variable whisker size
ggplot2::geom_pointrange(data = mod_table, ggplot2::aes(y = estimate,
x = name, ymin = lowerPR, ymax = upperPR, fill = name),
size = twig.size, fatten = trunk.size, shape = 23, stroke=2.2) + # Change point shape
scale_size_continuous(range = c(1, 14))+ # change point scaling
ggplot2::coord_flip() +
ggplot2::theme_bw() +
ggplot2::guides(fill = "none", colour = "none") +
ggplot2::theme(text=element_text(size=24, colour="black"))+ # Change font size
ggplot2::theme(legend.title = ggplot2::element_text(size = 15)) + # Increased font legend title
ggplot2::theme(legend.text = ggplot2::element_text(size = 13)) +
ggplot2::theme(legend.direction = "horizontal") +
ggplot2::theme(legend.background = ggplot2::element_blank()) +
ggplot2::labs(y = label, x = "", size = legend) +
ggplot2::theme(axis.text.y = ggplot2::element_text(size = 18,
colour = "black", hjust = 0.5, angle = angle)) + # Increased size title axis label
ggplot2::theme(axis.text.x = ggplot2::element_text(size = 18)) + # Increase size axis ticks
ggplot2::theme(panel.border = element_rect(colour = "black", fill=NA, size=1.3))
}
if (legend.pos == "bottom.right") {
plot <- plot + ggplot2::theme(legend.position = c(1,
0), legend.justification = c(1, 0))
}
else if (legend.pos == "bottom.left") {
plot <- plot + ggplot2::theme(legend.position = c(0,
0), legend.justification = c(0, 0))
}
else if (legend.pos == "top.right") {
plot <- plot + ggplot2::theme(legend.position = c(1,
1), legend.justification = c(1, 1))
}
else if (legend.pos == "top.left") {
plot <- plot + ggplot2::theme(legend.position = c(0,
1), legend.justification = c(0, 1))
}
else if (legend.pos == "top.out") {
plot <- plot + ggplot2::theme(legend.position = "top")
}
else if (legend.pos == "bottom.out") {
plot <- plot + ggplot2::theme(legend.position = "bottom")
}
if (cb == TRUE) {
plot <- plot + ggplot2::scale_fill_manual(values = cbpl) +
ggplot2::scale_colour_manual(values = cbpl)
}
if (k == TRUE && g == FALSE && k.pos == "right") {
plot <- plot + ggplot2::annotate("text", y = (max(data$yi) +
(max(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no]), # Size changed to 5.5
parse = TRUE, hjust = "right", size = 5.5)
}
else if (k == TRUE && g == FALSE && k.pos == "left") {
plot <- plot + ggplot2::annotate("text", y = (min(data$yi) +
(min(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no]),
parse = TRUE, hjust = "left", size = 5.5) # Size changed to 5.5
}
else if (k == TRUE && g == TRUE && k.pos == "right") {
plot <- plot + ggplot2::annotate("text", y = (max(data$yi) +
(max(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no],
" (", mod_table$g[1:group_no], ")"),
parse = TRUE, hjust = "right", size = 5.5) # Size changed to 5.5
}
else if (k == TRUE && g == TRUE && k.pos == "left") {
plot <- plot + ggplot2::annotate("text", y = (min(data$yi) +
(min(data$yi) * 0.1)), x = (seq(1, group_no, 1) +
0.3), label = paste("italic(k)==", mod_table$K[1:group_no],
" (", mod_table$g[1:group_no], ")"),
parse = TRUE, hjust = "left", size = 5.5) # Size changed to 5.5
}
return(plot)
}Intercept Meta-analytic model
int_model <- run.model(d, ~1)
summary(int_model)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -226.3154 452.6309 460.6309 480.5992 460.6678
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0095 0.0974 138 no species_ID no
## sigma^2.2 0.0249 0.1578 138 no phylogeny yes
## sigma^2.3 0.0605 0.2460 1089 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 1088) = 63442.0311, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.1898 0.0882 2.1525 137 0.0331 0.0154 0.3641 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(int_model, mod = "1", data = d, group = "species_ID") # For prediction intervals## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.1897603 0.01543568 0.3640848 -0.4437611 0.8232816i2_ml(int_model) # Estimate heterogeneity## I2_Total I2_species_ID I2_phylogeny I2_es_ID
## 99.481565 9.951012 26.095005 63.435548my.orchard(int_model, mod = "1", xlab = "dARR", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.04) # Display orchard plot
Single moderator metaregressions
Habitat
Individual coefficients
mod.habitat <- run.model(d, ~habitat - 1)
summary(mod.habitat)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -223.0240 446.0480 456.0480 481.0039 456.1035
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0099 0.0994 138 no species_ID no
## sigma^2.2 0.0134 0.1156 138 no phylogeny yes
## sigma^2.3 0.0606 0.2461 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56246.7341, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 136) = 7.2679, p-val = 0.0010
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## habitataquatic 0.2087 0.0664 3.1422 136 0.0021 0.0773 0.3400 **
## habitatterrestrial 0.0672 0.0817 0.8216 136 0.4127 -0.0945 0.2288
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.habitat, mod = "habitat", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Aquatic 0.20867269 0.07734442 0.3400010 -0.3786494 0.7959948
## 2 Terrestrial 0.06715867 -0.09448001 0.2287973 -0.5276751 0.6619924r2_ml(mod.habitat) # Calculate R2 (percentage of explained variation). Note that, throughout the manuscript, we report the marginal R2, which is the percentage of variation explained by the fixed effects## R2_marginal R2_conditional
## 0.02910927 0.29842002my.orchard(mod.habitat, mod = "habitat", xlab = "dARR", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.07)
Contrasts
mod.habitat_cont <- run.model(d, ~habitat) # Keep intercept for contrasts
summary(mod.habitat_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -223.0240 446.0480 456.0480 481.0039 456.1035
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0099 0.0994 138 no species_ID no
## sigma^2.2 0.0134 0.1156 138 no phylogeny yes
## sigma^2.3 0.0606 0.2461 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56246.7341, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 136) = 6.3394, p-val = 0.0130
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2087 0.0664 3.1422 136 0.0021 0.0773 0.3400
## habitatterrestrial -0.1415 0.0562 -2.5178 136 0.0130 -0.2527 -0.0304
##
## intrcpt **
## habitatterrestrial *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Heteroscedasticity check
Here, we can clearly see that the plastic responses of aquatic animals are more variable than the ones of terrestrial animals.
qplot(y = sqrt(residuals(mod.habitat)^2), x = fitted(mod.habitat)) + # plot sqrt(residuals^2) against fitted values
geom_point() +
geom_smooth(method = "lm") + # method ="lm" to generate a straight line
geom_hline(yintercept = 0, colour="red") # Seems heteroscedastic
Individual coefficients (with heteroscedasticity)
The rma.mv function has options to model heteroscedasticity. In our case, most of the heterogeneity is explained by residual variation (es_ID; within-species variation). Below, we model the heteroscedasticity at effect size level and compare AIC values to determine whether accounting for variance differences between habitats improves model’s fit.
##### Heteroscedasticity modeled at the effect size level
mod.habitat_het <- rma.mv(yi=dARR,
V=VCV_dARR, # run the model, as described earlier
mods= ~habitat-1,
method="REML",
test="t",
dfs="contain",
random=list(~1|species_ID,
~1|phylogeny,
~habitat|es_ID), # To model heteroscedasticity at the effect size level, for each habitat
struct="HCS", # heteroscedastic compound symmetric structure
rho=0, # Correlation of 0 for the HCS structure
R = list(phylogeny = phylo_matrix),
data=d,
sparse=TRUE)
AICc(mod.habitat)## [1] 456.1034AICc(mod.habitat_het) # Best fit## [1] 319.5823summary(mod.habitat_het)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -153.7523 307.5047 319.5047 349.4517 319.5824
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0948 138 no species_ID no
## sigma^2.2 0.0129 0.1136 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: habitat (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0716 0.2677 929 no aquatic
## tau^2.2 0.0082 0.0907 160 no terrestrial
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56246.7341, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 136) = 8.5055, p-val = 0.0003
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## habitataquatic 0.2087 0.0655 3.1856 136 0.0018 0.0792 0.3383 **
## habitatterrestrial 0.0595 0.0762 0.7814 136 0.4359 -0.0911 0.2102
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.habitat_het, mod="habitat",data=d, group="species_ID") ## name estimate lowerCL upperCL lowerPR upperPR
## 1 Aquatic 0.20872360 0.07915124 0.3382960 -0.4098660 0.8273132
## 2 Terrestrial 0.05952518 -0.09112232 0.2101727 -0.3152993 0.4343497r2_ml(mod.habitat_het)## R2_marginal R2_conditional
## 0.1130771 0.4772616my.orchard(mod.habitat_het, mod="habitat", xlab="dARR", alpha=0.1, data=d, group="species_ID", whisker=0.07)
Contrasts (with heteroscedasticity)
### Contrasts with heteroscedasticity
mod.habitat_het_cont <- rma.mv(yi=dARR,
V=VCV_dARR,
mods= ~habitat, # Keep intercept
method="REML",
test="t",
dfs="contain",
random=list(~1|species_ID,
~1|phylogeny,
~habitat|es_ID),
struct="HCS",
rho=0,
R = list(phylogeny = phylo_matrix),
data=d,
sparse=TRUE)
summary(mod.habitat_het_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -153.7523 307.5047 319.5047 349.4517 319.5824
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0948 138 no species_ID no
## sigma^2.2 0.0129 0.1136 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: habitat (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0716 0.2677 929 no aquatic
## tau^2.2 0.0082 0.0907 160 no terrestrial
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56246.7341, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 136) = 9.2383, p-val = 0.0028
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2087 0.0655 3.1856 136 0.0018 0.0792 0.3383
## habitatterrestrial -0.1492 0.0491 -3.0394 136 0.0028 -0.2463 -0.0521
##
## intrcpt **
## habitatterrestrial **
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Taxonomic groups
Individual coefficients
mod.taxa <- run.model(d, ~taxonomic_group - 1)
summary(mod.taxa)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -220.0038 440.0075 456.0075 495.9148 456.1415
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0957 138 no species_ID no
## sigma^2.2 0.0242 0.1554 138 no phylogeny yes
## sigma^2.3 0.0605 0.2459 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 1:5):
## F(df1 = 5, df2 = 133) = 2.0368, p-val = 0.0775
##
## Model Results:
##
## estimate se tval df pval
## taxonomic_groupamphibian 0.2083 0.1541 1.3521 133 0.1786
## taxonomic_groupaquatic invertebrate 0.1904 0.1131 1.6838 133 0.0946
## taxonomic_groupfish 0.2413 0.1161 2.0783 133 0.0396
## taxonomic_groupreptile 0.0667 0.1634 0.4082 133 0.6838
## taxonomic_groupterrestrial invertebrate 0.0535 0.1288 0.4157 133 0.6783
## ci.lb ci.ub
## taxonomic_groupamphibian -0.0964 0.5131
## taxonomic_groupaquatic invertebrate -0.0333 0.4142 .
## taxonomic_groupfish 0.0116 0.4709 *
## taxonomic_groupreptile -0.2564 0.3898
## taxonomic_groupterrestrial invertebrate -0.2012 0.3082
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.taxa, mod = "taxonomic_group", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR
## 1 Amphibian 0.20831727 -0.09642755 0.5130621 -0.4697867
## 2 Aquatic invertebrate 0.19043896 -0.03327497 0.4141529 -0.4553191
## 3 Fish 0.24127906 0.01164449 0.4709136 -0.4065539
## 4 Reptile 0.06668682 -0.25643776 0.3898114 -0.6198735
## 5 Terrestrial invertebrate 0.05352402 -0.20117964 0.3082277 -0.6036133
## upperPR
## 1 0.8864212
## 2 0.8361970
## 3 0.8891120
## 4 0.7532471
## 5 0.7106614r2_ml(mod.taxa)## R2_marginal R2_conditional
## 0.04420462 0.38371519my.orchard(mod.taxa, mod = "taxonomic_group", xlab = "dARR", alpha = 0.1, data = d,
group = "species_ID", whisker = 0.09)
Contrasts
# Reptiles as the reference group
mod.taxa_reptile_ref_cont <- run.model(d, ~relevel(factor(taxonomic_group), ref = "reptile"))
summary(mod.taxa_reptile_ref_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -220.0038 440.0075 456.0075 495.9148 456.1415
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0957 138 no species_ID no
## sigma^2.2 0.0242 0.1554 138 no phylogeny yes
## sigma^2.3 0.0605 0.2459 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3538, p-val = 0.2535
##
## Model Results:
##
## estimate
## intrcpt 0.0667
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.1416
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.1238
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.1746
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate -0.0132
## se
## intrcpt 0.1634
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.1070
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.1837
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.1550
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 0.1949
## tval
## intrcpt 0.4082
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 1.3231
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.6737
## relevel(factor(taxonomic_group), ref = "reptile")fish 1.1261
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate -0.0675
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 133
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "reptile")fish 133
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 133
## pval
## intrcpt 0.6838
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.1881
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.5016
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.2621
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 0.9463
## ci.lb
## intrcpt -0.2564
## relevel(factor(taxonomic_group), ref = "reptile")amphibian -0.0701
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate -0.2396
## relevel(factor(taxonomic_group), ref = "reptile")fish -0.1321
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate -0.3987
## ci.ub
## intrcpt 0.3898
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.3534
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.4871
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.4813
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 0.3724
##
## intrcpt
## relevel(factor(taxonomic_group), ref = "reptile")amphibian
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate
## relevel(factor(taxonomic_group), ref = "reptile")fish
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# Terrestrial invertebrates as the reference group
mod.taxa_tinvert_ref_cont <- run.model(d, ~relevel(factor(taxonomic_group), ref = "terrestrial invertebrate"))
summary(mod.taxa_tinvert_ref_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -220.0038 440.0075 456.0075 495.9148 456.1415
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0957 138 no species_ID no
## sigma^2.2 0.0242 0.1554 138 no phylogeny yes
## sigma^2.3 0.0605 0.2459 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3538, p-val = 0.2535
##
## Model Results:
##
## estimate
## intrcpt 0.0535
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.1548
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.1369
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.1878
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.0132
## se
## intrcpt 0.1288
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.1872
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.0768
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.1575
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.1949
## tval
## intrcpt 0.4157
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.8268
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 1.7824
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 1.1924
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.0675
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 133
## pval
## intrcpt 0.6783
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.4098
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.0770
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.2352
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.9463
## ci.lb
## intrcpt -0.2012
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian -0.2155
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate -0.0150
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish -0.1237
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile -0.3724
## ci.ub
## intrcpt 0.3082
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.5251
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.2888
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.4992
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.3987
##
## intrcpt
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate .
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# Aquatic invertebrates as the reference group
mod.taxa_ainvert_ref_cont <- run.model(d, ~relevel(factor(taxonomic_group), ref = "aquatic invertebrate"))
summary(mod.taxa_ainvert_ref_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -220.0038 440.0075 456.0075 495.9148 456.1415
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0957 138 no species_ID no
## sigma^2.2 0.0242 0.1554 138 no phylogeny yes
## sigma^2.3 0.0605 0.2459 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3538, p-val = 0.2535
##
## Model Results:
##
## estimate
## intrcpt 0.1904
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.0179
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.0508
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile -0.1238
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate -0.1369
## se
## intrcpt 0.1131
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.1755
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.1433
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 0.1837
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 0.0768
## tval
## intrcpt 1.6838
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.1019
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.3548
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile -0.6737
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate -1.7824
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 133
## pval
## intrcpt 0.0946
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.9190
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.7233
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 0.5016
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 0.0770
## ci.lb
## intrcpt -0.0333
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian -0.3292
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish -0.2326
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile -0.4871
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate -0.2888
## ci.ub
## intrcpt 0.4142
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.3649
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.3342
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 0.2396
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 0.0150
##
## intrcpt .
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# Amphibians as the reference group
mod.taxa_amph_ref_cont <- run.model(d, ~relevel(factor(taxonomic_group), ref = "amphibian"))
summary(mod.taxa_amph_ref_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -220.0038 440.0075 456.0075 495.9148 456.1415
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0957 138 no species_ID no
## sigma^2.2 0.0242 0.1554 138 no phylogeny yes
## sigma^2.3 0.0605 0.2459 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3538, p-val = 0.2535
##
## Model Results:
##
## estimate
## intrcpt 0.2083
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate -0.0179
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.0330
## relevel(factor(taxonomic_group), ref = "amphibian")reptile -0.1416
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate -0.1548
## se
## intrcpt 0.1541
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.1755
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.1452
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 0.1070
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 0.1872
## tval
## intrcpt 1.3521
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate -0.1019
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.2270
## relevel(factor(taxonomic_group), ref = "amphibian")reptile -1.3231
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate -0.8268
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "amphibian")fish 133
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 133
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 133
## pval
## intrcpt 0.1786
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.9190
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.8208
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 0.1881
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 0.4098
## ci.lb
## intrcpt -0.0964
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate -0.3649
## relevel(factor(taxonomic_group), ref = "amphibian")fish -0.2543
## relevel(factor(taxonomic_group), ref = "amphibian")reptile -0.3534
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate -0.5251
## ci.ub
## intrcpt 0.5131
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.3292
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.3202
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 0.0701
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 0.2155
##
## intrcpt
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate
## relevel(factor(taxonomic_group), ref = "amphibian")fish
## relevel(factor(taxonomic_group), ref = "amphibian")reptile
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1# Fishes as the reference group
mod.taxa_fish_ref_cont <- run.model(d, ~relevel(factor(taxonomic_group), ref = "fish"))
summary(mod.taxa_fish_ref_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -220.0038 440.0075 456.0075 495.9148 456.1415
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0957 138 no species_ID no
## sigma^2.2 0.0242 0.1554 138 no phylogeny yes
## sigma^2.3 0.0605 0.2459 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3538, p-val = 0.2535
##
## Model Results:
##
## estimate
## intrcpt 0.2413
## relevel(factor(taxonomic_group), ref = "fish")amphibian -0.0330
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate -0.0508
## relevel(factor(taxonomic_group), ref = "fish")reptile -0.1746
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate -0.1878
## se
## intrcpt 0.1161
## relevel(factor(taxonomic_group), ref = "fish")amphibian 0.1452
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 0.1433
## relevel(factor(taxonomic_group), ref = "fish")reptile 0.1550
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 0.1575
## tval
## intrcpt 2.0783
## relevel(factor(taxonomic_group), ref = "fish")amphibian -0.2270
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate -0.3548
## relevel(factor(taxonomic_group), ref = "fish")reptile -1.1261
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate -1.1924
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "fish")amphibian 133
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "fish")reptile 133
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 133
## pval
## intrcpt 0.0396
## relevel(factor(taxonomic_group), ref = "fish")amphibian 0.8208
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 0.7233
## relevel(factor(taxonomic_group), ref = "fish")reptile 0.2621
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 0.2352
## ci.lb
## intrcpt 0.0116
## relevel(factor(taxonomic_group), ref = "fish")amphibian -0.3202
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate -0.3342
## relevel(factor(taxonomic_group), ref = "fish")reptile -0.4813
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate -0.4992
## ci.ub
## intrcpt 0.4709
## relevel(factor(taxonomic_group), ref = "fish")amphibian 0.2543
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 0.2326
## relevel(factor(taxonomic_group), ref = "fish")reptile 0.1321
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 0.1237
##
## intrcpt *
## relevel(factor(taxonomic_group), ref = "fish")amphibian
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate
## relevel(factor(taxonomic_group), ref = "fish")reptile
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Heteroscedasticity check
qplot(y = sqrt(residuals(mod.taxa)^2), x = fitted(mod.taxa)) + geom_point() + geom_smooth(method = "lm") +
geom_hline(yintercept = 0, colour = "red") # Seems heteroscedastic
Individual coefficients (with heteroscedasticity)
Here, the heteroscedasticity is also modeled at both the effect size and cohort levels, which improved the model fit.
mod.taxa_het <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~taxonomic_group - 1, method = "REML",
test = "t", dfs = "contain", random = list(~1 | species_ID, ~1 | phylogeny, ~taxonomic_group |
es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix), data = d,
sparse = TRUE)
AICc(mod.taxa)## [1] 456.1409AICc(mod.taxa_het) # Improved model fit## [1] 133.3114summary(mod.taxa_het)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -54.5107 109.0214 133.0214 192.8824 133.3128
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0946 138 no species_ID no
## sigma^2.2 0.0296 0.1722 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.1878 0.4334 71 no amphibian
## tau^2.2 0.1357 0.3684 221 no aquatic invertebrate
## tau^2.3 0.0329 0.1813 623 no fish
## tau^2.4 0.0163 0.1278 27 no reptile
## tau^2.5 0.0069 0.0833 147 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 1:5):
## F(df1 = 5, df2 = 133) = 1.9108, p-val = 0.0966
##
## Model Results:
##
## estimate se tval df pval
## taxonomic_groupamphibian 0.1966 0.1761 1.1167 133 0.2661
## taxonomic_groupaquatic invertebrate 0.1992 0.1286 1.5492 133 0.1237
## taxonomic_groupfish 0.2538 0.1265 2.0058 133 0.0469
## taxonomic_groupreptile 0.0702 0.1733 0.4053 133 0.6859
## taxonomic_groupterrestrial invertebrate 0.0488 0.1410 0.3464 133 0.7296
## ci.lb ci.ub
## taxonomic_groupamphibian -0.1517 0.5449
## taxonomic_groupaquatic invertebrate -0.0551 0.4535
## taxonomic_groupfish 0.0035 0.5041 *
## taxonomic_groupreptile -0.2726 0.4131
## taxonomic_groupterrestrial invertebrate -0.2301 0.3278
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.taxa_het, mod = "taxonomic_group", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR
## 1 Amphibian 0.19664022 -0.151655428 0.5449359 -0.8069271
## 2 Aquatic invertebrate 0.19919762 -0.055126478 0.4535217 -0.6649603
## 3 Fish 0.25382273 0.003518141 0.5041273 -0.3312188
## 4 Reptile 0.07024502 -0.272560393 0.4130504 -0.5063220
## 5 Terrestrial invertebrate 0.04884628 -0.230059070 0.3277516 -0.4571120
## upperPR
## 1 1.2002076
## 2 1.0633556
## 3 0.8388642
## 4 0.6468121
## 5 0.5548046r2_ml(mod.taxa_het)## R2_marginal R2_conditional
## 0.1170636 0.3219044my.orchard(mod.taxa_het, mod = "taxonomic_group", xlab = "dARR", alpha = 0.1, data = d,
group = "species_ID", whisker = 0.09)
Contrasts (with heteroscedasticity)
## Reptiles as the reference group
mod.taxa_het_cont_rept_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(taxonomic_group),
ref = "reptile"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~taxonomic_group | es_ID), struct = "HCS", rho = 0,
R = list(phylogeny = phylo_matrix), data = d, sparse = TRUE)
summary(mod.taxa_het_cont_rept_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -54.5107 109.0214 133.0214 192.8824 133.3128
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0946 138 no species_ID no
## sigma^2.2 0.0296 0.1722 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.1878 0.4334 71 no amphibian
## tau^2.2 0.1357 0.3684 221 no aquatic invertebrate
## tau^2.3 0.0329 0.1813 623 no fish
## tau^2.4 0.0163 0.1278 27 no reptile
## tau^2.5 0.0069 0.0833 147 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3982, p-val = 0.2380
##
## Model Results:
##
## estimate
## intrcpt 0.0702
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.1264
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.1290
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.1836
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate -0.0214
## se
## intrcpt 0.1733
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.1140
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.2002
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.1615
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 0.2098
## tval
## intrcpt 0.4053
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 1.1090
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.6442
## relevel(factor(taxonomic_group), ref = "reptile")fish 1.1365
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate -0.1020
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 133
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "reptile")fish 133
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 133
## pval
## intrcpt 0.6859
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.2694
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.5206
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.2578
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 0.9189
## ci.lb
## intrcpt -0.2726
## relevel(factor(taxonomic_group), ref = "reptile")amphibian -0.0990
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate -0.2670
## relevel(factor(taxonomic_group), ref = "reptile")fish -0.1359
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate -0.4363
## ci.ub
## intrcpt 0.4131
## relevel(factor(taxonomic_group), ref = "reptile")amphibian 0.3518
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate 0.5249
## relevel(factor(taxonomic_group), ref = "reptile")fish 0.5031
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate 0.3935
##
## intrcpt
## relevel(factor(taxonomic_group), ref = "reptile")amphibian
## relevel(factor(taxonomic_group), ref = "reptile")aquatic invertebrate
## relevel(factor(taxonomic_group), ref = "reptile")fish
## relevel(factor(taxonomic_group), ref = "reptile")terrestrial invertebrate
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Terrestrial invertebrates as the reference group
mod.taxa_het_cont_tinvert_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(taxonomic_group),
ref = "terrestrial invertebrate"), method = "REML", test = "t", dfs = "contain",
random = list(~1 | species_ID, ~1 | phylogeny, ~taxonomic_group | es_ID), struct = "HCS",
rho = 0, R = list(phylogeny = phylo_matrix), data = d, sparse = TRUE)
summary(mod.taxa_het_cont_tinvert_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -54.5107 109.0214 133.0214 192.8824 133.3128
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0946 138 no species_ID no
## sigma^2.2 0.0296 0.1722 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.1878 0.4334 71 no amphibian
## tau^2.2 0.1357 0.3684 221 no aquatic invertebrate
## tau^2.3 0.0329 0.1813 623 no fish
## tau^2.4 0.0163 0.1278 27 no reptile
## tau^2.5 0.0069 0.0833 147 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3982, p-val = 0.2380
##
## Model Results:
##
## estimate
## intrcpt 0.0488
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.1478
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.1504
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.2050
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.0214
## se
## intrcpt 0.1410
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.2121
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.0794
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.1731
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.2098
## tval
## intrcpt 0.3464
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.6969
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 1.8938
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 1.1839
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.1020
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 133
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 133
## pval
## intrcpt 0.7296
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.4871
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.0604
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.2386
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.9189
## ci.lb
## intrcpt -0.2301
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian -0.2717
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate -0.0067
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish -0.1375
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile -0.3935
## ci.ub
## intrcpt 0.3278
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian 0.5672
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate 0.3074
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish 0.5474
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile 0.4363
##
## intrcpt
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")amphibian
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")aquatic invertebrate .
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")fish
## relevel(factor(taxonomic_group), ref = "terrestrial invertebrate")reptile
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Aquatic invertebrates as the reference group
mod.taxa_het_cont_ainvert_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(taxonomic_group),
ref = "aquatic invertebrate"), method = "REML", test = "t", dfs = "contain",
random = list(~1 | species_ID, ~1 | phylogeny, ~taxonomic_group | es_ID), struct = "HCS",
rho = 0, R = list(phylogeny = phylo_matrix), data = d, sparse = TRUE)
summary(mod.taxa_het_cont_ainvert_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -54.5107 109.0214 133.0214 192.8824 133.3128
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0946 138 no species_ID no
## sigma^2.2 0.0296 0.1722 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.1878 0.4334 71 no amphibian
## tau^2.2 0.1357 0.3684 221 no aquatic invertebrate
## tau^2.3 0.0329 0.1813 623 no fish
## tau^2.4 0.0163 0.1278 27 no reptile
## tau^2.5 0.0069 0.0833 147 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3982, p-val = 0.2380
##
## Model Results:
##
## estimate
## intrcpt 0.1992
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian -0.0026
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.0546
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile -0.1290
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate -0.1504
## se
## intrcpt 0.1286
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.2026
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.1614
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 0.2002
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 0.0794
## tval
## intrcpt 1.5492
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian -0.0126
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.3385
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile -0.6442
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate -1.8938
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 133
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 133
## pval
## intrcpt 0.1237
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.9899
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.7355
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 0.5206
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 0.0604
## ci.lb
## intrcpt -0.0551
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian -0.4033
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish -0.2646
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile -0.5249
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate -0.3074
## ci.ub
## intrcpt 0.4535
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian 0.3981
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish 0.3738
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile 0.2670
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate 0.0067
##
## intrcpt
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")amphibian
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")fish
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")reptile
## relevel(factor(taxonomic_group), ref = "aquatic invertebrate")terrestrial invertebrate .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Amphibians as the reference group
mod.taxa_het_cont_amph_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(taxonomic_group),
ref = "amphibian"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~taxonomic_group | es_ID), struct = "HCS", rho = 0,
R = list(phylogeny = phylo_matrix), data = d, sparse = TRUE)
summary(mod.taxa_het_cont_amph_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -54.5107 109.0214 133.0214 192.8824 133.3128
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0946 138 no species_ID no
## sigma^2.2 0.0296 0.1722 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.1878 0.4334 71 no amphibian
## tau^2.2 0.1357 0.3684 221 no aquatic invertebrate
## tau^2.3 0.0329 0.1813 623 no fish
## tau^2.4 0.0163 0.1278 27 no reptile
## tau^2.5 0.0069 0.0833 147 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3982, p-val = 0.2380
##
## Model Results:
##
## estimate
## intrcpt 0.1966
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.0026
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.0572
## relevel(factor(taxonomic_group), ref = "amphibian")reptile -0.1264
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate -0.1478
## se
## intrcpt 0.1761
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.2026
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.1645
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 0.1140
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 0.2121
## tval
## intrcpt 1.1167
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.0126
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.3476
## relevel(factor(taxonomic_group), ref = "amphibian")reptile -1.1090
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate -0.6969
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "amphibian")fish 133
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 133
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 133
## pval
## intrcpt 0.2661
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.9899
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.7287
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 0.2694
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 0.4871
## ci.lb
## intrcpt -0.1517
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate -0.3981
## relevel(factor(taxonomic_group), ref = "amphibian")fish -0.2682
## relevel(factor(taxonomic_group), ref = "amphibian")reptile -0.3518
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate -0.5672
## ci.ub
## intrcpt 0.5449
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate 0.4033
## relevel(factor(taxonomic_group), ref = "amphibian")fish 0.3826
## relevel(factor(taxonomic_group), ref = "amphibian")reptile 0.0990
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate 0.2717
##
## intrcpt
## relevel(factor(taxonomic_group), ref = "amphibian")aquatic invertebrate
## relevel(factor(taxonomic_group), ref = "amphibian")fish
## relevel(factor(taxonomic_group), ref = "amphibian")reptile
## relevel(factor(taxonomic_group), ref = "amphibian")terrestrial invertebrate
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Fishes as the reference group
mod.taxa_het_cont_fish_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(taxonomic_group),
ref = "fish"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~taxonomic_group | es_ID), struct = "HCS", rho = 0,
R = list(phylogeny = phylo_matrix), data = d, sparse = TRUE)
summary(mod.taxa_het_cont_fish_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -54.5107 109.0214 133.0214 192.8824 133.3128
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0946 138 no species_ID no
## sigma^2.2 0.0296 0.1722 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.1878 0.4334 71 no amphibian
## tau^2.2 0.1357 0.3684 221 no aquatic invertebrate
## tau^2.3 0.0329 0.1813 623 no fish
## tau^2.4 0.0163 0.1278 27 no reptile
## tau^2.5 0.0069 0.0833 147 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 2:5):
## F(df1 = 4, df2 = 133) = 1.3982, p-val = 0.2380
##
## Model Results:
##
## estimate
## intrcpt 0.2538
## relevel(factor(taxonomic_group), ref = "fish")amphibian -0.0572
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate -0.0546
## relevel(factor(taxonomic_group), ref = "fish")reptile -0.1836
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate -0.2050
## se
## intrcpt 0.1265
## relevel(factor(taxonomic_group), ref = "fish")amphibian 0.1645
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 0.1614
## relevel(factor(taxonomic_group), ref = "fish")reptile 0.1615
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 0.1731
## tval
## intrcpt 2.0058
## relevel(factor(taxonomic_group), ref = "fish")amphibian -0.3476
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate -0.3385
## relevel(factor(taxonomic_group), ref = "fish")reptile -1.1365
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate -1.1839
## df
## intrcpt 133
## relevel(factor(taxonomic_group), ref = "fish")amphibian 133
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 133
## relevel(factor(taxonomic_group), ref = "fish")reptile 133
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 133
## pval
## intrcpt 0.0469
## relevel(factor(taxonomic_group), ref = "fish")amphibian 0.7287
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 0.7355
## relevel(factor(taxonomic_group), ref = "fish")reptile 0.2578
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 0.2386
## ci.lb
## intrcpt 0.0035
## relevel(factor(taxonomic_group), ref = "fish")amphibian -0.3826
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate -0.3738
## relevel(factor(taxonomic_group), ref = "fish")reptile -0.5031
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate -0.5474
## ci.ub
## intrcpt 0.5041
## relevel(factor(taxonomic_group), ref = "fish")amphibian 0.2682
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate 0.2646
## relevel(factor(taxonomic_group), ref = "fish")reptile 0.1359
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate 0.1375
##
## intrcpt *
## relevel(factor(taxonomic_group), ref = "fish")amphibian
## relevel(factor(taxonomic_group), ref = "fish")aquatic invertebrate
## relevel(factor(taxonomic_group), ref = "fish")reptile
## relevel(factor(taxonomic_group), ref = "fish")terrestrial invertebrate
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Persistence of developmental effects
Here the effect sizes estimating the “direct” and “persistent” effects of developmental temperatures on heat tolerance are compared.
Individual coefficients
mod.persist <- run.model(d, ~brought_common_temp - 1)
summary(mod.persist)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -212.9028 425.8057 435.8057 460.7616 435.8612
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0800 138 no species_ID no
## sigma^2.2 0.0142 0.1193 138 no phylogeny yes
## sigma^2.3 0.0602 0.2454 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56507.5086, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 1087) = 19.2928, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## brought_common_tempno 0.2214 0.0678 3.2640 1087 0.0011 0.0883 0.3544
## brought_common_tempyes 0.0394 0.0728 0.5419 1087 0.5880 -0.1033 0.1822
##
## brought_common_tempno **
## brought_common_tempyes
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persist, mod = "brought_common_temp", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 No 0.22135272 0.08828745 0.3544180 -0.3522484 0.7949539
## 2 Yes 0.03943152 -0.10334461 0.1822077 -0.5364998 0.6153629r2_ml(mod.persist)## R2_marginal R2_conditional
## 0.06253941 0.30188023my.orchard(mod.persist, mod = "brought_common_temp", xlab = "dARR", alpha = 0.1,
data = d, group = "species_ID", whisker = 0.07)
Contrasts
mod.persist_cont <- run.model(d, ~brought_common_temp)
summary(mod.persist_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -212.9028 425.8057 435.8057 460.7616 435.8612
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0800 138 no species_ID no
## sigma^2.2 0.0142 0.1193 138 no phylogeny yes
## sigma^2.3 0.0602 0.2454 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56507.5086, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 30.7459, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb
## intrcpt 0.2214 0.0678 3.2640 136 0.0014 0.0872
## brought_common_tempyes -0.1819 0.0328 -5.5449 1087 <.0001 -0.2463
## ci.ub
## intrcpt 0.3555 **
## brought_common_tempyes -0.1175 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Heteroscedasticity check
qplot(y = sqrt(residuals(mod.persist)^2), x = fitted(mod.persist)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red") # Seems fine
Individual coefficients (with heteroscedasticity)
mod.persist_het <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~brought_common_temp -
1, method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~brought_common_temp | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
AICc(mod.persist)## [1] 435.8611AICc(mod.persist_het) # Improved model fit## [1] 380.4215summary(mod.persist_het)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -184.1719 368.3439 380.3439 410.2909 380.4216
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0765 138 no species_ID no
## sigma^2.2 0.0155 0.1245 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: brought_common_temp (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0696 0.2639 866 no no
## tau^2.2 0.0210 0.1450 223 no yes
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56507.5086, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 1087) = 22.8750, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## brought_common_tempno 0.2244 0.0705 3.1811 1087 0.0015 0.0860 0.3628
## brought_common_tempyes 0.0485 0.0733 0.6626 1087 0.5078 -0.0952 0.1923
##
## brought_common_tempno **
## brought_common_tempyes
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persist_het, mod = "brought_common_temp", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 No 0.22439541 0.08598476 0.3628061 -0.3834860 0.8322768
## 2 Yes 0.04853997 -0.09521006 0.1922900 -0.3802659 0.4773458r2_ml(mod.persist_het)## R2_marginal R2_conditional
## 0.1908900 0.4127618my.orchard(mod.persist_het, mod = "brought_common_temp", xlab = "dARR", alpha = 0.1,
data = d, group = "species_ID", whisker = 0.07)
Contrasts (with heteroscedasticity)
mod.persist_het_cont <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~brought_common_temp,
method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~brought_common_temp | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.persist_het_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -184.1719 368.3439 380.3439 410.2909 380.4216
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0765 138 no species_ID no
## sigma^2.2 0.0155 0.1245 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: brought_common_temp (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0696 0.2639 866 no no
## tau^2.2 0.0210 0.1450 223 no yes
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56507.5086, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 39.1689, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb
## intrcpt 0.2244 0.0705 3.1811 136 0.0018 0.0849
## brought_common_tempyes -0.1759 0.0281 -6.2585 1087 <.0001 -0.2310
## ci.ub
## intrcpt 0.3639 **
## brought_common_tempyes -0.1207 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Life history variation and persistence
Here, the types of experimental designs presented in Figure 2 are compared to estimate the life history variation and persistence of developmental effects.
Individual coefficients
mod.design <- run.model(d, ~exp_design - 1)
summary(mod.design)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 1:6):
## F(df1 = 6, df2 = 1083) = 9.0167, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## exp_designA 0.2219 0.0676 3.2805 1083 0.0011 0.0892 0.3547 **
## exp_designB 0.1170 0.0964 1.2139 1083 0.2250 -0.0721 0.3061
## exp_designC 0.2503 0.0756 3.3123 1083 0.0010 0.1020 0.3986 ***
## exp_designD -0.0930 0.0821 -1.1319 1083 0.2579 -0.2541 0.0682
## exp_designE 0.1057 0.0832 1.2695 1083 0.2045 -0.0577 0.2690
## exp_designF 0.1131 0.0815 1.3880 1083 0.1654 -0.0468 0.2731
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.design, mod = "exp_design", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 A 0.22192299 0.08918504 0.35466095 -0.3474271 0.7912731
## 2 B 0.11700190 -0.07211415 0.30611795 -0.4680665 0.7020703
## 3 C 0.25030917 0.10202997 0.39858836 -0.3228635 0.8234818
## 4 D -0.09295895 -0.25410017 0.06818226 -0.6695928 0.4836749
## 5 E 0.10567936 -0.05765618 0.26901490 -0.4715716 0.6829303
## 6 F 0.11313922 -0.04680240 0.27308084 -0.4631606 0.6894390r2_ml(mod.design)## R2_marginal R2_conditional
## 0.09918076 0.32410191my.orchard(mod.design, mod = "exp_design", xlab = "dARR", alpha = 0.1, data = d,
group = "species_ID", whisker = 0.09)
Contrasts
## Design A as the reference group
mod.design_cont_A_ref <- run.model(d, ~relevel(factor(exp_design), ref = "A"))
summary(mod.design_cont_A_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 9.2212, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.2219 0.0676 3.2805 132
## relevel(factor(exp_design), ref = "A")B -0.1049 0.0726 -1.4458 1083
## relevel(factor(exp_design), ref = "A")C 0.0284 0.0421 0.6750 1083
## relevel(factor(exp_design), ref = "A")D -0.3149 0.0513 -6.1329 1083
## relevel(factor(exp_design), ref = "A")E -0.1162 0.0530 -2.1921 1083
## relevel(factor(exp_design), ref = "A")F -0.1088 0.0503 -2.1613 1083
## pval ci.lb ci.ub
## intrcpt 0.0013 0.0881 0.3557 **
## relevel(factor(exp_design), ref = "A")B 0.1485 -0.2473 0.0375
## relevel(factor(exp_design), ref = "A")C 0.4998 -0.0541 0.1109
## relevel(factor(exp_design), ref = "A")D <.0001 -0.4156 -0.2141 ***
## relevel(factor(exp_design), ref = "A")E 0.0286 -0.2203 -0.0122 *
## relevel(factor(exp_design), ref = "A")F 0.0309 -0.2075 -0.0100 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design B as the reference group
mod.design_cont_B_ref <- run.model(d, ~relevel(factor(exp_design), ref = "B"))
summary(mod.design_cont_B_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 9.2212, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.1170 0.0964 1.2139 132
## relevel(factor(exp_design), ref = "B")A 0.1049 0.0726 1.4458 1083
## relevel(factor(exp_design), ref = "B")C 0.1333 0.0789 1.6892 1083
## relevel(factor(exp_design), ref = "B")D -0.2100 0.0872 -2.4089 1083
## relevel(factor(exp_design), ref = "B")E -0.0113 0.0885 -0.1279 1083
## relevel(factor(exp_design), ref = "B")F -0.0039 0.0862 -0.0448 1083
## pval ci.lb ci.ub
## intrcpt 0.2269 -0.0737 0.3077
## relevel(factor(exp_design), ref = "B")A 0.1485 -0.0375 0.2473
## relevel(factor(exp_design), ref = "B")C 0.0915 -0.0215 0.2882 .
## relevel(factor(exp_design), ref = "B")D 0.0162 -0.3810 -0.0389 *
## relevel(factor(exp_design), ref = "B")E 0.8982 -0.1850 0.1624
## relevel(factor(exp_design), ref = "B")F 0.9643 -0.1730 0.1653
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design C as the reference group
mod.design_cont_C_ref <- run.model(d, ~relevel(factor(exp_design), ref = "C"))
summary(mod.design_cont_C_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 9.2212, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.2503 0.0756 3.3123 132
## relevel(factor(exp_design), ref = "C")A -0.0284 0.0421 -0.6750 1083
## relevel(factor(exp_design), ref = "C")B -0.1333 0.0789 -1.6892 1083
## relevel(factor(exp_design), ref = "C")D -0.3433 0.0635 -5.4077 1083
## relevel(factor(exp_design), ref = "C")E -0.1446 0.0631 -2.2925 1083
## relevel(factor(exp_design), ref = "C")F -0.1372 0.0619 -2.2147 1083
## pval ci.lb ci.ub
## intrcpt 0.0012 0.1008 0.3998 **
## relevel(factor(exp_design), ref = "C")A 0.4998 -0.1109 0.0541
## relevel(factor(exp_design), ref = "C")B 0.0915 -0.2882 0.0215 .
## relevel(factor(exp_design), ref = "C")D <.0001 -0.4678 -0.2187 ***
## relevel(factor(exp_design), ref = "C")E 0.0221 -0.2684 -0.0208 *
## relevel(factor(exp_design), ref = "C")F 0.0270 -0.2587 -0.0156 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design D as the reference group
mod.design_cont_D_ref <- run.model(d, ~relevel(factor(exp_design), ref = "D"))
summary(mod.design_cont_D_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 9.2212, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt -0.0930 0.0821 -1.1319 132
## relevel(factor(exp_design), ref = "D")A 0.3149 0.0513 6.1329 1083
## relevel(factor(exp_design), ref = "D")B 0.2100 0.0872 2.4089 1083
## relevel(factor(exp_design), ref = "D")C 0.3433 0.0635 5.4077 1083
## relevel(factor(exp_design), ref = "D")E 0.1986 0.0683 2.9104 1083
## relevel(factor(exp_design), ref = "D")F 0.2061 0.0697 2.9567 1083
## pval ci.lb ci.ub
## intrcpt 0.2597 -0.2554 0.0695
## relevel(factor(exp_design), ref = "D")A <.0001 0.2141 0.4156 ***
## relevel(factor(exp_design), ref = "D")B 0.0162 0.0389 0.3810 *
## relevel(factor(exp_design), ref = "D")C <.0001 0.2187 0.4678 ***
## relevel(factor(exp_design), ref = "D")E 0.0037 0.0647 0.3326 **
## relevel(factor(exp_design), ref = "D")F 0.0032 0.0693 0.3429 **
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design E as the reference group
mod.design_cont_E_ref <- run.model(d, ~relevel(factor(exp_design), ref = "E"))
summary(mod.design_cont_E_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 9.2212, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.1057 0.0832 1.2695 132
## relevel(factor(exp_design), ref = "E")A 0.1162 0.0530 2.1921 1083
## relevel(factor(exp_design), ref = "E")B 0.0113 0.0885 0.1279 1083
## relevel(factor(exp_design), ref = "E")C 0.1446 0.0631 2.2925 1083
## relevel(factor(exp_design), ref = "E")D -0.1986 0.0683 -2.9104 1083
## relevel(factor(exp_design), ref = "E")F 0.0075 0.0683 0.1093 1083
## pval ci.lb ci.ub
## intrcpt 0.2065 -0.0590 0.2703
## relevel(factor(exp_design), ref = "E")A 0.0286 0.0122 0.2203 *
## relevel(factor(exp_design), ref = "E")B 0.8982 -0.1624 0.1850
## relevel(factor(exp_design), ref = "E")C 0.0221 0.0208 0.2684 *
## relevel(factor(exp_design), ref = "E")D 0.0037 -0.3326 -0.0647 **
## relevel(factor(exp_design), ref = "E")F 0.9130 -0.1265 0.1414
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design F as the reference group
mod.design_cont_F_ref <- run.model(d, ~relevel(factor(exp_design), ref = "F"))
summary(mod.design_cont_F_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 9.2212, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.1131 0.0815 1.3880 132
## relevel(factor(exp_design), ref = "F")A 0.1088 0.0503 2.1613 1083
## relevel(factor(exp_design), ref = "F")B 0.0039 0.0862 0.0448 1083
## relevel(factor(exp_design), ref = "F")C 0.1372 0.0619 2.2147 1083
## relevel(factor(exp_design), ref = "F")D -0.2061 0.0697 -2.9567 1083
## relevel(factor(exp_design), ref = "F")E -0.0075 0.0683 -0.1093 1083
## pval ci.lb ci.ub
## intrcpt 0.1675 -0.0481 0.2744
## relevel(factor(exp_design), ref = "F")A 0.0309 0.0100 0.2075 *
## relevel(factor(exp_design), ref = "F")B 0.9643 -0.1653 0.1730
## relevel(factor(exp_design), ref = "F")C 0.0270 0.0156 0.2587 *
## relevel(factor(exp_design), ref = "F")D 0.0032 -0.3429 -0.0693 **
## relevel(factor(exp_design), ref = "F")E 0.9130 -0.1414 0.1265
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Heteroscedasticity check
qplot(y = sqrt(residuals(mod.design)^2), x = fitted(mod.design)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Individual coefficients (with heteroscedasticity)
mod.design_het <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~exp_design - 1, method = "REML",
test = "t", dfs = "contain", random = list(~1 | species_ID, ~1 | phylogeny, ~exp_design |
es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix), data = d,
sparse = TRUE)
AICc(mod.design)## [1] 430.0181AICc(mod.design_het) # Better fit## [1] 274.205summary(mod.design_het)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: exp_design (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0753 0.2745 700 no A
## tau^2.2 0.3209 0.5665 20 no B
## tau^2.3 0.0155 0.1244 146 no C
## tau^2.4 0.0352 0.1876 92 no D
## tau^2.5 0.0134 0.1156 76 no E
## tau^2.6 0.0074 0.0859 55 no F
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 1:6):
## F(df1 = 6, df2 = 1083) = 10.4670, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## exp_designA 0.2302 0.0740 3.1108 1083 0.0019 0.0850 0.3755 **
## exp_designB 0.0981 0.1570 0.6249 1083 0.5321 -0.2099 0.4062
## exp_designC 0.2502 0.0782 3.2004 1083 0.0014 0.0968 0.4036 **
## exp_designD -0.0818 0.0849 -0.9630 1083 0.3358 -0.2483 0.0848
## exp_designE 0.1181 0.0825 1.4316 1083 0.1525 -0.0438 0.2800
## exp_designF 0.1015 0.0784 1.2956 1083 0.1954 -0.0522 0.2553
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.design_het, mod = "exp_design", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 A 0.23024864 0.08501963 0.37547765 -0.4032517 0.8637490
## 2 B 0.09811306 -0.20993386 0.40615999 -1.0937752 1.2900013
## 3 C 0.25017802 0.09679244 0.40356360 -0.1661248 0.6664808
## 4 D -0.08176060 -0.24834760 0.08482639 -0.5851651 0.4216439
## 5 E 0.11811684 -0.04377422 0.28000790 -0.2915607 0.5277944
## 6 F 0.10153337 -0.05223750 0.25530425 -0.2756146 0.4786813r2_ml(mod.design_het)## R2_marginal R2_conditional
## 0.2705497 0.4689134my.orchard(mod.design_het, mod = "exp_design", xlab = "dARR", alpha = 0.1, data = d,
group = "species_ID", whisker = 0.09)
Contrasts (with heteroscedasticity)
## Design A as the reference group
mod.design_het_cont_A_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(exp_design),
ref = "A"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~exp_design | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.design_het_cont_A_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: exp_design (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0753 0.2745 700 no A
## tau^2.2 0.3209 0.5665 20 no B
## tau^2.3 0.0155 0.1244 146 no C
## tau^2.4 0.0352 0.1876 92 no D
## tau^2.5 0.0134 0.1156 76 no E
## tau^2.6 0.0074 0.0859 55 no F
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 11.2372, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.2302 0.0740 3.1108 132
## relevel(factor(exp_design), ref = "A")B -0.1321 0.1406 -0.9397 1083
## relevel(factor(exp_design), ref = "A")C 0.0199 0.0368 0.5412 1083
## relevel(factor(exp_design), ref = "A")D -0.3120 0.0477 -6.5421 1083
## relevel(factor(exp_design), ref = "A")E -0.1121 0.0429 -2.6142 1083
## relevel(factor(exp_design), ref = "A")F -0.1287 0.0343 -3.7561 1083
## pval ci.lb ci.ub
## intrcpt 0.0023 0.0838 0.3767 **
## relevel(factor(exp_design), ref = "A")B 0.3476 -0.4081 0.1438
## relevel(factor(exp_design), ref = "A")C 0.5885 -0.0523 0.0922
## relevel(factor(exp_design), ref = "A")D <.0001 -0.4056 -0.2184 ***
## relevel(factor(exp_design), ref = "A")E 0.0091 -0.1963 -0.0280 **
## relevel(factor(exp_design), ref = "A")F 0.0002 -0.1960 -0.0615 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design B as the reference group
mod.design_het_cont_B_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(exp_design),
ref = "B"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~exp_design | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.design_het_cont_B_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: exp_design (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0753 0.2745 700 no A
## tau^2.2 0.3209 0.5665 20 no B
## tau^2.3 0.0155 0.1244 146 no C
## tau^2.4 0.0352 0.1876 92 no D
## tau^2.5 0.0134 0.1156 76 no E
## tau^2.6 0.0074 0.0859 55 no F
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 11.2372, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.0981 0.1570 0.6249 132
## relevel(factor(exp_design), ref = "B")A 0.1321 0.1406 0.9397 1083
## relevel(factor(exp_design), ref = "B")C 0.1521 0.1420 1.0710 1083
## relevel(factor(exp_design), ref = "B")D -0.1799 0.1473 -1.2214 1083
## relevel(factor(exp_design), ref = "B")E 0.0200 0.1458 0.1372 1083
## relevel(factor(exp_design), ref = "B")F 0.0034 0.1427 0.0240 1083
## pval ci.lb ci.ub
## intrcpt 0.5331 -0.2124 0.4087
## relevel(factor(exp_design), ref = "B")A 0.3476 -0.1438 0.4081
## relevel(factor(exp_design), ref = "B")C 0.2844 -0.1265 0.4307
## relevel(factor(exp_design), ref = "B")D 0.2222 -0.4688 0.1091
## relevel(factor(exp_design), ref = "B")E 0.8909 -0.2661 0.3061
## relevel(factor(exp_design), ref = "B")F 0.9809 -0.2765 0.2834
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design C as the reference group
mod.design_het_cont_C_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(exp_design),
ref = "C"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~exp_design | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.design_het_cont_C_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: exp_design (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0753 0.2745 700 no A
## tau^2.2 0.3209 0.5665 20 no B
## tau^2.3 0.0155 0.1244 146 no C
## tau^2.4 0.0352 0.1876 92 no D
## tau^2.5 0.0134 0.1156 76 no E
## tau^2.6 0.0074 0.0859 55 no F
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 11.2372, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.2502 0.0782 3.2004 132
## relevel(factor(exp_design), ref = "C")A -0.0199 0.0368 -0.5412 1083
## relevel(factor(exp_design), ref = "C")B -0.1521 0.1420 -1.0710 1083
## relevel(factor(exp_design), ref = "C")D -0.3319 0.0562 -5.9095 1083
## relevel(factor(exp_design), ref = "C")E -0.1321 0.0512 -2.5801 1083
## relevel(factor(exp_design), ref = "C")F -0.1486 0.0450 -3.3017 1083
## pval ci.lb ci.ub
## intrcpt 0.0017 0.0955 0.4048 **
## relevel(factor(exp_design), ref = "C")A 0.5885 -0.0922 0.0523
## relevel(factor(exp_design), ref = "C")B 0.2844 -0.4307 0.1265
## relevel(factor(exp_design), ref = "C")D <.0001 -0.4422 -0.2217 ***
## relevel(factor(exp_design), ref = "C")E 0.0100 -0.2325 -0.0316 *
## relevel(factor(exp_design), ref = "C")F 0.0010 -0.2370 -0.0603 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design D as the reference group
mod.design_het_cont_D_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(exp_design),
ref = "D"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~exp_design | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.design_het_cont_D_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: exp_design (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0753 0.2745 700 no A
## tau^2.2 0.3209 0.5665 20 no B
## tau^2.3 0.0155 0.1244 146 no C
## tau^2.4 0.0352 0.1876 92 no D
## tau^2.5 0.0134 0.1156 76 no E
## tau^2.6 0.0074 0.0859 55 no F
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 11.2372, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt -0.0818 0.0849 -0.9630 132
## relevel(factor(exp_design), ref = "D")A 0.3120 0.0477 6.5421 1083
## relevel(factor(exp_design), ref = "D")B 0.1799 0.1473 1.2214 1083
## relevel(factor(exp_design), ref = "D")C 0.3319 0.0562 5.9095 1083
## relevel(factor(exp_design), ref = "D")E 0.1999 0.0535 3.7378 1083
## relevel(factor(exp_design), ref = "D")F 0.1833 0.0550 3.3308 1083
## pval ci.lb ci.ub
## intrcpt 0.3373 -0.2497 0.0862
## relevel(factor(exp_design), ref = "D")A <.0001 0.2184 0.4056 ***
## relevel(factor(exp_design), ref = "D")B 0.2222 -0.1091 0.4688
## relevel(factor(exp_design), ref = "D")C <.0001 0.2217 0.4422 ***
## relevel(factor(exp_design), ref = "D")E 0.0002 0.0950 0.3048 ***
## relevel(factor(exp_design), ref = "D")F 0.0009 0.0753 0.2913 ***
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design E as the reference group
mod.design_het_cont_E_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(exp_design),
ref = "E"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~exp_design | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.design_het_cont_E_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: exp_design (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0753 0.2745 700 no A
## tau^2.2 0.3209 0.5665 20 no B
## tau^2.3 0.0155 0.1244 146 no C
## tau^2.4 0.0352 0.1876 92 no D
## tau^2.5 0.0134 0.1156 76 no E
## tau^2.6 0.0074 0.0859 55 no F
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 11.2372, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.1181 0.0825 1.4316 132
## relevel(factor(exp_design), ref = "E")A 0.1121 0.0429 2.6142 1083
## relevel(factor(exp_design), ref = "E")B -0.0200 0.1458 -0.1372 1083
## relevel(factor(exp_design), ref = "E")C 0.1321 0.0512 2.5801 1083
## relevel(factor(exp_design), ref = "E")D -0.1999 0.0535 -3.7378 1083
## relevel(factor(exp_design), ref = "E")F -0.0166 0.0485 -0.3418 1083
## pval ci.lb ci.ub
## intrcpt 0.1546 -0.0451 0.2813
## relevel(factor(exp_design), ref = "E")A 0.0091 0.0280 0.1963 **
## relevel(factor(exp_design), ref = "E")B 0.8909 -0.3061 0.2661
## relevel(factor(exp_design), ref = "E")C 0.0100 0.0316 0.2325 *
## relevel(factor(exp_design), ref = "E")D 0.0002 -0.3048 -0.0950 ***
## relevel(factor(exp_design), ref = "E")F 0.7326 -0.1118 0.0786
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Design F as the reference group
mod.design_het_cont_F_ref <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~relevel(factor(exp_design),
ref = "F"), method = "REML", test = "t", dfs = "contain", random = list(~1 |
species_ID, ~1 | phylogeny, ~exp_design | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.design_het_cont_F_ref)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: exp_design (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0753 0.2745 700 no A
## tau^2.2 0.3209 0.5665 20 no B
## tau^2.3 0.0155 0.1244 146 no C
## tau^2.4 0.0352 0.1876 92 no D
## tau^2.5 0.0134 0.1156 76 no E
## tau^2.6 0.0074 0.0859 55 no F
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 11.2372, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.1015 0.0784 1.2956 132
## relevel(factor(exp_design), ref = "F")A 0.1287 0.0343 3.7561 1083
## relevel(factor(exp_design), ref = "F")B -0.0034 0.1427 -0.0240 1083
## relevel(factor(exp_design), ref = "F")C 0.1486 0.0450 3.3017 1083
## relevel(factor(exp_design), ref = "F")D -0.1833 0.0550 -3.3308 1083
## relevel(factor(exp_design), ref = "F")E 0.0166 0.0485 0.3418 1083
## pval ci.lb ci.ub
## intrcpt 0.1974 -0.0535 0.2566
## relevel(factor(exp_design), ref = "F")A 0.0002 0.0615 0.1960 ***
## relevel(factor(exp_design), ref = "F")B 0.9809 -0.2834 0.2765
## relevel(factor(exp_design), ref = "F")C 0.0010 0.0603 0.2370 ***
## relevel(factor(exp_design), ref = "F")D 0.0009 -0.2913 -0.0753 ***
## relevel(factor(exp_design), ref = "F")E 0.7326 -0.0786 0.1118
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Time in a common garden condition
Run model and plot results
mod.common_temp <- run.model(d, ~scale(time_common_temp, center = T, scale = F)) # Center to zero mean
summary(mod.common_temp)##
## Multivariate Meta-Analysis Model (k = 204; method: REML)
##
## logLik Deviance AIC BIC AICc
## 24.3962 -48.7923 -38.7923 -22.2510 -38.4862
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0038 0.0618 21 no species_ID no
## sigma^2.2 0.0516 0.2273 21 no phylogeny yes
## sigma^2.3 0.0203 0.1425 204 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 202) = 7121.3816, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 202) = 8.9912, p-val = 0.0031
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.0677 0.1324 0.5115 19
## scale(time_common_temp, center = T, scale = F) 0.0014 0.0005 2.9985 202
## pval ci.lb ci.ub
## intrcpt 0.6149 -0.2094 0.3449
## scale(time_common_temp, center = T, scale = F) 0.0031 0.0005 0.0024 **
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(mod.common_temp)## R2_marginal R2_conditional
## 0.04787401 0.74470467d.common_temp <- filter(d, time_common_temp != "NA")
plot_continuous(d.common_temp, mod.common_temp, d.common_temp$time_common_temp, "Time at common temperature (days)")
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.common_temp)^2), x = fitted(mod.common_temp)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red") # Seems relatively stable
Type of metric
Individual coefficients
mod.metric <- run.model(d, ~metric - 1)
summary(mod.metric)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -225.7190 451.4380 461.4380 486.3939 461.4935
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0096 0.0979 138 no species_ID no
## sigma^2.2 0.0257 0.1603 138 no phylogeny yes
## sigma^2.3 0.0604 0.2458 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 63440.4166, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 1087) = 2.9226, p-val = 0.0542
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## metricCTmax 0.1967 0.0897 2.1930 1087 0.0285 0.0207 0.3727 *
## metricLT50 0.1646 0.0921 1.7868 1087 0.0743 -0.0162 0.3453 .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.metric, mod = "metric", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 CTmax 0.1966785 0.02070034 0.3726566 -0.4352999 0.8286569
## 2 LT50 0.1645513 -0.01615094 0.3452536 -0.4687588 0.7978615r2_ml(mod.metric)## R2_marginal R2_conditional
## 0.001772338 0.369636784my.orchard(mod.metric, mod = "metric", xlab = "dARR", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.07)
Contrasts
mod.metric_cont <- run.model(d, ~metric)
summary(mod.metric_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -225.7190 451.4380 461.4380 486.3939 461.4935
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0096 0.0979 138 no species_ID no
## sigma^2.2 0.0257 0.1603 138 no phylogeny yes
## sigma^2.3 0.0604 0.2458 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 63440.4166, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 1.3473, p-val = 0.2460
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.1967 0.0897 2.1930 136 0.0300 0.0193 0.3740 *
## metricLT50 -0.0321 0.0277 -1.1608 1087 0.2460 -0.0864 0.0222
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Heteroscedasticity check
qplot(y = sqrt(residuals(mod.metric)^2), x = fitted(mod.metric)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red") # Seems fine
Individual coefficients (with heteroscedasticity)
mod.metric_het <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~metric - 1, method = "REML",
test = "t", dfs = "contain", random = list(~1 | species_ID, ~1 | phylogeny, ~metric |
es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix), data = d,
sparse = TRUE)
AICc(mod.metric)## [1] 461.4934AICc(mod.metric_het) # Better fit## [1] 454.4918summary(mod.metric_het)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -221.2071 442.4142 454.4142 484.3612 454.4920
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0095 0.0976 138 no species_ID no
## sigma^2.2 0.0263 0.1622 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: metric (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0645 0.2539 863 no CTmax
## tau^2.2 0.0409 0.2022 226 no LT50
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 63440.4166, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 1087) = 2.9452, p-val = 0.0530
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## metricCTmax 0.1945 0.0907 2.1453 1087 0.0321 0.0166 0.3724 *
## metricLT50 0.1616 0.0925 1.7470 1087 0.0809 -0.0199 0.3430 .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.metric_het, mod = "metric", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 CTmax 0.1945252 0.01660836 0.3724420 -0.4519017 0.8409521
## 2 LT50 0.1615711 -0.01989991 0.3430421 -0.4114067 0.7345489r2_ml(mod.metric_het)## R2_marginal R2_conditional
## 0.004964216 0.269282977my.orchard(mod.metric_het, mod = "metric", xlab = "dARR", alpha = 0.1, data = d,
group = "species_ID", whisker = 0.07)
Contrasts (with heteroscedasticity)
mod.metric_het_cont <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~metric, method = "REML",
test = "t", dfs = "contain", random = list(~1 | species_ID, ~1 | phylogeny, ~metric |
es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix), data = d,
sparse = TRUE)
summary(mod.metric_het_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -221.2071 442.4142 454.4142 484.3612 454.4920
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0095 0.0976 138 no species_ID no
## sigma^2.2 0.0263 0.1622 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: metric (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0645 0.2539 863 no CTmax
## tau^2.2 0.0409 0.2022 226 no LT50
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 63440.4166, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 1.6464, p-val = 0.1997
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.1945 0.0907 2.1453 136 0.0337 0.0152 0.3738 *
## metricLT50 -0.0330 0.0257 -1.2831 1087 0.1997 -0.0833 0.0174
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Heating rate
Run model and plot results
mod.ramping <- run.model(d, ~scale(ramping, center = T, scale = F)) # Center to zero mean
summary(mod.ramping)##
## Multivariate Meta-Analysis Model (k = 855; method: REML)
##
## logLik Deviance AIC BIC AICc
## -165.8218 331.6436 341.6436 365.3874 341.7144
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0072 0.0847 114 no species_ID no
## sigma^2.2 0.0347 0.1863 114 no phylogeny yes
## sigma^2.3 0.0638 0.2526 855 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 853) = 55693.6919, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 853) = 0.3524, p-val = 0.5529
##
## Model Results:
##
## estimate se tval df pval
## intrcpt 0.2116 0.1045 2.0241 112 0.0453
## scale(ramping, center = T, scale = F) 0.0194 0.0327 0.5937 853 0.5529
## ci.lb ci.ub
## intrcpt 0.0045 0.4187 *
## scale(ramping, center = T, scale = F) -0.0448 0.0836
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(mod.ramping)## R2_marginal R2_conditional
## 0.000681348 0.396621099d.ramping <- filter(d, ramping != "NA")
plot_continuous(d.ramping, mod.ramping, d.ramping$ramping, "Heating rate (degrees/min)")
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.ramping)^2), x = fitted(mod.ramping)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Mean adjusted for habitat differences
Overall mean with habitat weighted equally
mod.adjusted_mean <- mod_results(model = mod.habitat_het, data = d, mod = "1", weights = "equal",
group = "species_ID")
mod.adjusted_mean$mod_table## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.1341244 0.002268286 0.2659805 -0.4547758 0.7230246my.orchard(mod.adjusted_mean, mod = "1", xlab = "dARR", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.04)
Multi-model inference
We used the MuMin package to generate all combinations of moderators, and determine the set of models explaining the most variation
eval(metafor:::.MuMIn) # Required for the models to run.
full_model.MuMIn<- rma.mv(dARR, V=VCV_dARR,
method="ML", # maximum likelihood for model selection
test="t",
dfs="contain",
mods= ~ habitat + # All moderators without missing values
taxonomic_group +
metric +
exp_design,
random=list(~1|species_ID,
~1|phylogeny,
~1|es_ID),
R = list(phylogeny = phylo_matrix),
data=d,
sparse=TRUE)
options(na.action = "na.fail") #required for dredge to run
candidate_models<-dredge(full_model.MuMIn) # Generate all possible combinations of moderators
options(na.action = "na.omit") #set back to default
subset(candidate_models, delta<=2) # Display all models within 2 values of AICc## Global model call: rma.mv(yi = dARR, V = VCV_dARR, mods = ~habitat + taxonomic_group +
## metric + exp_design, random = list(~1 | species_ID, ~1 |
## phylogeny, ~1 | es_ID), data = d, method = "ML", test = "t",
## dfs = "contain", R = list(phylogeny = phylo_matrix), sparse = TRUE)
## ---
## Model selection table
## (Int) exp_dsg hbt mtr txn_grp df logLik AICc delta weight
## 8 + + + + 11 -204.748 431.7 0.00 0.334
## 4 + + + 10 -206.212 432.6 0.89 0.214
## 6 + + + 10 -206.531 433.3 1.53 0.156
## 2 + + 9 -207.569 433.3 1.56 0.153
## 14 + + + + 14 -202.516 433.4 1.68 0.144
## Models ranked by AICc(x)sw(model.avg(candidate_models, subset=delta<=2))# relative importance (sum of weights) of the moderators## exp_design metric habitat taxonomic_group
## Sum of weights: 1.00 0.63 0.55 0.14
## N containing models: 5 3 2 1Best models
Best model
best_model<- run.model(d, ~ habitat + # Paste best model
metric +
exp_design)
summary(best_model) # summary of the best model identified ##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -204.1003 408.2006 430.2006 485.0427 430.4476
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0063 0.0793 138 no species_ID no
## sigma^2.2 0.0098 0.0992 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1081) = 45713.3143, p-val < .0001
##
## Test of Moderators (coefficients 2:8):
## F(df1 = 7, df2 = 130) = 7.2357, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2369 0.0583 4.0654 130 <.0001 0.1216 0.3522
## habitatterrestrial -0.0632 0.0532 -1.1866 130 0.2375 -0.1685 0.0422
## metricLT50 -0.0416 0.0275 -1.5106 1081 0.1312 -0.0956 0.0124
## exp_designB -0.0820 0.0744 -1.1022 1081 0.2706 -0.2280 0.0640
## exp_designC 0.0317 0.0422 0.7526 1081 0.4519 -0.0510 0.1145
## exp_designD -0.3035 0.0533 -5.6932 1081 <.0001 -0.4081 -0.1989
## exp_designE -0.1114 0.0544 -2.0486 1081 0.0407 -0.2182 -0.0047
## exp_designF -0.1153 0.0505 -2.2839 1081 0.0226 -0.2144 -0.0162
##
## intrcpt ***
## habitatterrestrial
## metricLT50
## exp_designB
## exp_designC
## exp_designD ***
## exp_designE *
## exp_designF *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(best_model) # R-squared of the best model ## R2_marginal R2_conditional
## 0.1043967 0.2947820# Assumptions
qplot(y = sqrt(residuals(best_model)^2), x = fitted(best_model)) +
geom_point() +
geom_smooth(method = "lm") +
geom_hline(yintercept = 0, colour="red") 
Second best model
best_model_n2 <- run.model(d, ~habitat + exp_design)
summary(best_model_n2)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.1119 410.2238 430.2238 480.0894 430.4292
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0062 0.0785 138 no species_ID no
## sigma^2.2 0.0104 0.1022 138 no phylogeny yes
## sigma^2.3 0.0598 0.2446 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1082) = 45719.9218, p-val < .0001
##
## Test of Moderators (coefficients 2:7):
## F(df1 = 6, df2 = 131) = 8.0454, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2283 0.0595 3.8344 131 0.0002 0.1105 0.3461
## habitatterrestrial -0.0574 0.0535 -1.0740 131 0.2848 -0.1632 0.0483
## exp_designB -0.1068 0.0725 -1.4731 1082 0.1410 -0.2491 0.0355
## exp_designC 0.0275 0.0420 0.6535 1082 0.5136 -0.0550 0.1099
## exp_designD -0.3010 0.0532 -5.6557 1082 <.0001 -0.4055 -0.1966
## exp_designE -0.1058 0.0542 -1.9507 1082 0.0514 -0.2122 0.0006
## exp_designF -0.1078 0.0503 -2.1443 1082 0.0322 -0.2064 -0.0092
##
## intrcpt ***
## habitatterrestrial
## exp_designB
## exp_designC
## exp_designD ***
## exp_designE .
## exp_designF *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(best_model_n2)## R2_marginal R2_conditional
## 0.1030582 0.2979860# Assumptions
qplot(y = sqrt(residuals(best_model_n2)^2), x = fitted(best_model_n2)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Third best model
best_model_n3 <- run.model(d, ~metric + exp_design)
summary(best_model_n3)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.0087 410.0174 430.0174 479.8831 430.2229
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0144 0.1201 138 no phylogeny yes
## sigma^2.3 0.0597 0.2442 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1082) = 47853.3762, p-val < .0001
##
## Test of Moderators (coefficients 2:7):
## F(df1 = 6, df2 = 1082) = 8.0079, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2295 0.0688 3.3358 131 0.0011 0.0934 0.3656 **
## metricLT50 -0.0398 0.0275 -1.4456 1082 0.1486 -0.0939 0.0142
## exp_designB -0.0808 0.0744 -1.0852 1082 0.2781 -0.2268 0.0653
## exp_designC 0.0331 0.0422 0.7856 1082 0.4323 -0.0496 0.1159
## exp_designD -0.3180 0.0514 -6.1822 1082 <.0001 -0.4190 -0.2171 ***
## exp_designE -0.1218 0.0533 -2.2876 1082 0.0224 -0.2263 -0.0173 *
## exp_designF -0.1162 0.0506 -2.2957 1082 0.0219 -0.2154 -0.0169 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(best_model_n3)## R2_marginal R2_conditional
## 0.09859645 0.32754534# Assumptions
qplot(y = sqrt(residuals(best_model_n3)^2), x = fitted(best_model_n3)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Fourth best model
best_model_n4 <- run.model(d, ~exp_design)
summary(best_model_n4)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -205.9256 411.8513 429.8513 474.7387 430.0190
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0059 0.0768 138 no species_ID no
## sigma^2.2 0.0140 0.1182 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 2:6):
## F(df1 = 5, df2 = 1083) = 9.2212, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2219 0.0676 3.2805 132 0.0013 0.0881 0.3557 **
## exp_designB -0.1049 0.0726 -1.4458 1083 0.1485 -0.2473 0.0375
## exp_designC 0.0284 0.0421 0.6750 1083 0.4998 -0.0541 0.1109
## exp_designD -0.3149 0.0513 -6.1329 1083 <.0001 -0.4156 -0.2141 ***
## exp_designE -0.1162 0.0530 -2.1921 1083 0.0286 -0.2203 -0.0122 *
## exp_designF -0.1088 0.0503 -2.1613 1083 0.0309 -0.2075 -0.0100 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(best_model_n4)## R2_marginal R2_conditional
## 0.09918076 0.32410191# Assumptions
qplot(y = sqrt(residuals(best_model_n4)^2), x = fitted(best_model_n4)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Fifth best model
best_model_n5 <- run.model(d, ~exp_design + metric + taxonomic_group)
summary(best_model_n5)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -200.4149 400.8297 428.8297 498.5898 429.2248
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0053 0.0730 138 no species_ID no
## sigma^2.2 0.0165 0.1283 138 no phylogeny yes
## sigma^2.3 0.0597 0.2444 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1078) = 43868.7717, p-val < .0001
##
## Test of Moderators (coefficients 2:11):
## F(df1 = 10, df2 = 127) = 5.2608, p-val < .0001
##
## Model Results:
##
## estimate se tval df
## intrcpt 0.1951 0.1293 1.5084 127
## exp_designB -0.0785 0.0744 -1.0548 1078
## exp_designC 0.0390 0.0420 0.9279 1078
## exp_designD -0.3164 0.0527 -6.0025 1078
## exp_designE -0.0976 0.0541 -1.8050 1078
## exp_designF -0.1097 0.0507 -2.1612 1078
## metricLT50 -0.0428 0.0276 -1.5502 1078
## taxonomic_groupaquatic invertebrate 0.0411 0.1494 0.2752 127
## taxonomic_groupfish 0.0688 0.1222 0.5626 127
## taxonomic_groupreptile 0.0196 0.0980 0.1998 127
## taxonomic_groupterrestrial invertebrate -0.0836 0.1591 -0.5253 127
## pval ci.lb ci.ub
## intrcpt 0.1339 -0.0608 0.4510
## exp_designB 0.2917 -0.2246 0.0675
## exp_designC 0.3537 -0.0435 0.1214
## exp_designD <.0001 -0.4199 -0.2130 ***
## exp_designE 0.0714 -0.2038 0.0085 .
## exp_designF 0.0309 -0.2092 -0.0101 *
## metricLT50 0.1214 -0.0970 0.0114
## taxonomic_groupaquatic invertebrate 0.7836 -0.2545 0.3367
## taxonomic_groupfish 0.5747 -0.1731 0.3106
## taxonomic_groupreptile 0.8419 -0.1744 0.2136
## taxonomic_groupterrestrial invertebrate 0.6003 -0.3985 0.2313
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(best_model_n5)## R2_marginal R2_conditional
## 0.1120662 0.3493867# Assumptions
qplot(y = sqrt(residuals(best_model_n5)^2), x = fitted(best_model_n5)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Publication bias
Funnel plot
funnel(int_model,
yaxis="seinv", # Inverse of standard error (precision) as the y axis
level = c(90, 95), # levels of statistical significance highlighted
shade = c("white", "gray55"), # shades for different levels of statistical significance
legend = TRUE, # display legend
ylab="Precision (1/SE)",
cex.lab=1.5,
digits=1,
xlim=c(-1.8,1.8),
col=ifelse(d$imputed=="no", "blue", "red")) #imputed values in red, original values in blue
Data type
Whether the data was published (reported in the publication) or unpublished (provided by authors)
Run model and plot results
mod.data_type <- run.model(d, ~data_type - 1)
summary(mod.data_type)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -226.2466 452.4933 462.4933 487.4492 462.5488
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0096 0.0978 138 no species_ID no
## sigma^2.2 0.0241 0.1553 138 no phylogeny yes
## sigma^2.3 0.0606 0.2461 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 58602.4405, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 1087) = 2.4945, p-val = 0.0830
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## data_typeobtained 0.1692 0.0973 1.7380 1087 0.0825 -0.0218 0.3602 .
## data_typepublished 0.1912 0.0869 2.1997 1087 0.0280 0.0206 0.3617 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1my.orchard(mod.data_type, mod = "data_type", xlab = "dARR", alpha = 0.1, data = d,
group = "species_ID", whisker = 0.07)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.data_type)^2), x = fitted(mod.data_type)) + # plot sqrt(residuals^2) against fitted values
geom_point() +
geom_smooth(method = "lm") + # method ="lm" to generate a straight line
geom_hline(yintercept = 0, colour="red") 
Contrasts
mod.data_type_cont <- run.model(d, ~data_type)
summary(mod.data_type_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -226.2466 452.4933 462.4933 487.4492 462.5488
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0096 0.0978 138 no species_ID no
## sigma^2.2 0.0241 0.1553 138 no phylogeny yes
## sigma^2.3 0.0606 0.2461 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 58602.4405, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 0.2176, p-val = 0.6409
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.1692 0.0973 1.7380 136 0.0845 -0.0233 0.3617 .
## data_typepublished 0.0220 0.0471 0.4665 1087 0.6409 -0.0705 0.1144
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Publication type
Whether the data was peer-reviewed or not-peer-reviewed (PhD or MSc dissertations)
Run model and plot results
mod.pub_type <- run.model(d, ~peer.reviewed - 1)
summary(mod.pub_type)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -225.1086 450.2173 460.2173 485.1731 460.2728
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0960 138 no species_ID no
## sigma^2.2 0.0234 0.1530 138 no phylogeny yes
## sigma^2.3 0.0606 0.2462 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 57727.1219, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 1087) = 3.5577, p-val = 0.0288
##
## Model Results:
##
## estimate se tval df pval ci.lb
## peer.reviewednot-peer-reviewed 0.1332 0.0937 1.4214 1087 0.1555 -0.0507
## peer.reviewedpeer-reviewed 0.1962 0.0857 2.2886 1087 0.0223 0.0280
## ci.ub
## peer.reviewednot-peer-reviewed 0.3170
## peer.reviewedpeer-reviewed 0.3645 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1my.orchard(mod.pub_type, mod = "peer.reviewed", xlab = "dARR", alpha = 0.1, data = d,
group = "species_ID", whisker = 0.07)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.pub_type)^2), x = fitted(mod.pub_type)) + # plot sqrt(residuals^2) against fitted values
geom_point() +
geom_smooth(method = "lm") + # method ="lm" to generate a straight line
geom_hline(yintercept = 0, colour="red") 
Contrasts
mod.pub_type_cont <- run.model(d, ~peer.reviewed)
summary(mod.pub_type_cont)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -225.1086 450.2173 460.2173 485.1731 460.2728
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0092 0.0960 138 no species_ID no
## sigma^2.2 0.0234 0.1530 138 no phylogeny yes
## sigma^2.3 0.0606 0.2462 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 57727.1219, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 2.2070, p-val = 0.1377
##
## Model Results:
##
## estimate se tval df pval ci.lb
## intrcpt 0.1332 0.0937 1.4214 136 0.1575 -0.0521
## peer.reviewedpeer-reviewed 0.0630 0.0424 1.4856 1087 0.1377 -0.0202
## ci.ub
## intrcpt 0.3185
## peer.reviewedpeer-reviewed 0.1463
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Publication year
Test for time-lag bias
Run model and plot results
mod.pub_year <- run.model(d, ~scale(pub_year, center = T, scale = F)) # Center to zero mean
summary(mod.pub_year)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -225.9969 451.9939 461.9939 486.9497 462.0494
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0095 0.0976 138 no species_ID no
## sigma^2.2 0.0250 0.1581 138 no phylogeny yes
## sigma^2.3 0.0605 0.2460 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 61209.7562, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 0.1047, p-val = 0.7463
##
## Model Results:
##
## estimate se tval df pval
## intrcpt 0.1913 0.0885 2.1621 136 0.0324
## scale(pub_year, center = T, scale = F) -0.0004 0.0013 -0.3236 1087 0.7463
## ci.lb ci.ub
## intrcpt 0.0163 0.3663 *
## scale(pub_year, center = T, scale = F) -0.0029 0.0021
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot_continuous(d, mod.pub_year, d$pub_year, "Publication year")
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.pub_year)^2), x = fitted(mod.pub_year)) + # plot sqrt(residuals^2) against fitted values
geom_point() +
geom_smooth(method = "lm") + # method ="lm" to generate a straight line
geom_hline(yintercept = 0, colour="red") 
Egger’s regression
We performed an Egger’s regression by fitting standard error (sqrt(Var_dARR)) as a moderator
mod.Egger <- run.model(d, ~sqrt(Var_dARR)) # standard error = sqrt(Var_dARRD)
summary(mod.Egger)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -224.7914 449.5829 459.5829 484.5387 459.6384
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0096 0.0979 138 no species_ID no
## sigma^2.2 0.0248 0.1574 138 no phylogeny yes
## sigma^2.3 0.0605 0.2460 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 63292.8875, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 0.0853, p-val = 0.7703
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.1930 0.0887 2.1759 136 0.0313 0.0176 0.3685 *
## sqrt(Var_dARR) -0.0396 0.1356 -0.2921 1087 0.7703 -0.3056 0.2264
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1pred<-predict.rma(mod.Egger)
d %>% mutate(fit=pred$pred,
ci.lb=pred$ci.lb,
ci.ub=pred$ci.ub,
pr.lb=pred$cr.lb,
pr.ub=pred$cr.ub) %>%
ggplot(aes(x = sqrt(Var_dARR), y = dARR)) +
geom_ribbon(aes(ymin = pr.lb, ymax = pr.ub, color = NULL), alpha = .1) +
geom_ribbon(aes(ymin = ci.lb, ymax = ci.ub, color = NULL), alpha = .3) +
geom_point(size=2,shape=21, alpha=0.7, fill="sienna1", col="gray25",stroke=1) +
geom_line(aes(y = fit), size = 1.2)+
labs(x = "Standard error", y = "dARR") +
theme_bw() +
geom_hline(yintercept = 0,linetype = 2, colour = "black",alpha=0.5)+ # horizontal line at lnRR = 0
theme(text = element_text(size = 18, colour = "black", hjust = 0.5), # change font sizes and legend position
legend.text=element_text(size=14),
legend.position=c(0,0),
legend.justification = c(0,0),
legend.background = element_blank(),
legend.direction="horizontal",
legend.title = element_text(size=15),
panel.border=element_rect(colour="black", fill=NA, size=1.2))
Sampling variance fitted as a moderator
mod.var <- run.model(d, ~Var_dARR)
summary(mod.var)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -223.6832 447.3664 457.3664 482.3223 457.4219
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0095 0.0975 138 no species_ID no
## sigma^2.2 0.0247 0.1572 138 no phylogeny yes
## sigma^2.3 0.0605 0.2460 1089 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 63441.4554, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 1087) = 0.1562, p-val = 0.6928
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.1903 0.0879 2.1653 136 0.0321 0.0165 0.3641 *
## Var_dARR -0.0341 0.0863 -0.3952 1087 0.6928 -0.2034 0.1352
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1pred<-predict.rma(mod.var)
d %>% mutate(fit=pred$pred,
ci.lb=pred$ci.lb,
ci.ub=pred$ci.ub,
pr.lb=pred$cr.lb,
pr.ub=pred$cr.ub) %>%
ggplot(aes(x = Var_dARR, y = dARR)) +
geom_ribbon(aes(ymin = pr.lb, ymax = pr.ub, color = NULL), alpha = .1) +
geom_ribbon(aes(ymin = ci.lb, ymax = ci.ub, color = NULL), alpha = .3) +
geom_point(size=2,shape=21, alpha=0.7, fill="sienna1", col="gray25",stroke=1) +
geom_line(aes(y = fit), size = 1.2)+
labs(x = "Sampling variance", y = "dARR") +
theme_bw() +
geom_hline(yintercept = 0,linetype = 2, colour = "black",alpha=0.5)+ # horizontal line at lnRR = 0
theme(text = element_text(size = 18, colour = "black", hjust = 0.5), # change font sizes and legend position
legend.text=element_text(size=14),
legend.position=c(0,0),
legend.justification = c(0,0),
legend.background = element_blank(),
legend.direction="horizontal",
legend.title = element_text(size=15),
panel.border=element_rect(colour="black", fill=NA, size=1.2))
Sensitivity analyses
Leave-one-out analyses
Here, we iteratively removed one study or one species at a time and investigated how it affected the overall mean. We also tried to iteratively remove one independent comparison (shared_trt_ID) because it was too computationally extensive.
Leave one study out
d$study_ID<-as.factor(d$study_ID)
d<-as.data.frame(d) # Only work with a dataframe
VCV_matrix<-list() # will need new VCV matrices because the sample size will be iteratively reduced
Leave1studyout<-list() # create a list that will host the results of each model
for(i in 1:length(levels(d$study_ID))){ # N models = N studies
VCV_matrix[[i]]<-make_VCV_matrix(d[d$study_ID != levels(d$study_ID)[i], ], V="Var_dARR", cluster="shared_trt_ID", obs="es_ID") # Create a new VCV matrix for each new model
Leave1studyout[[i]] <- rma.mv(yi = dARR, V = VCV_matrix[[i]], # Same model structure as all the models we fitted
method="REML",
test="t",
dfs="contain",
random = list(~1|species_ID,
~1|phylogeny,
~1|es_ID),
R= list(phylogeny = phylo_matrix),
data = d[d$study_ID != levels(d$study_ID)[i], ],
sparse=TRUE) # Generate a new model for each new data (iterative removal of one study at a time)
}
# The output is a list so we need to summarise the coefficients of all the models performed
results.Leave1studyout<-as.data.frame(cbind(
sapply(Leave1studyout, function(x) summary(x)$beta), # extract the beta coefficient from all models
sapply(Leave1studyout, function(x) summary(x)$se), # extract the standard error from all models
sapply(Leave1studyout, function(x) summary(x)$zval), # extract the z value from all models
sapply(Leave1studyout, function(x) summary(x)$pval), # extract the p value from all models
sapply(Leave1studyout, function(x) summary(x)$ci.lb), # extract the lower confidence interval for all models
sapply(Leave1studyout, function(x) summary(x)$ci.ub))) # extract the upper confidence interval for all models
colnames(results.Leave1studyout)=c("Estimate", "SE", "zval", "pval", "ci.lb", "ci.ub") # change column names
kable(results.Leave1studyout)%>% kable_styling("striped", position="left") %>% scroll_box(width="100%", height="500px") # Table of the results from all models
ggplot(results.Leave1studyout)+stat_dots(aes(x=Estimate), alpha=0.8, dotsize=1.5, shape=16, show.legend=F, col="darkcyan")
results.Leave1studyout %>%
summarise(estimate= mean(Estimate), # calculate the mean coefficients across the models generated
se=mean(SE),
zval=mean(zval),
pval=mean(pval),
ci.lb=mean(ci.lb),
ci.ub=mean(ci.ub)) %>%
kable()%>% kable_styling("striped", position="left")load(here("Rdata", "Leaveonestudyout.RData")) # Load the output of the models to avoid running them
kable(results.Leave1studyout)%>% kable_styling("striped", position="left") %>% scroll_box(width="100%", height="500px") # Table of the results from all models| Estimate | SE | zval | pval | ci.lb | ci.ub |
|---|---|---|---|---|---|
| 0.1891962 | 0.0888518 | 2.129345 | 0.0350140 | 0.0134978 | 0.3648945 |
| 0.1896830 | 0.0881920 | 2.150796 | 0.0332471 | 0.0152894 | 0.3640766 |
| 0.1935150 | 0.0922495 | 2.097736 | 0.0377789 | 0.0110861 | 0.3759439 |
| 0.1893077 | 0.0890196 | 2.126585 | 0.0352472 | 0.0132776 | 0.3653377 |
| 0.1901109 | 0.0868494 | 2.188972 | 0.0303053 | 0.0183609 | 0.3618609 |
| 0.1898975 | 0.0887416 | 2.139893 | 0.0341484 | 0.0144056 | 0.3653894 |
| 0.1901953 | 0.0921256 | 2.064522 | 0.0408684 | 0.0080114 | 0.3723792 |
| 0.1897198 | 0.0879738 | 2.156548 | 0.0327865 | 0.0157576 | 0.3636819 |
| 0.1917912 | 0.0921903 | 2.080384 | 0.0393950 | 0.0094549 | 0.3741275 |
| 0.1895508 | 0.0891313 | 2.126646 | 0.0352420 | 0.0132997 | 0.3658019 |
| 0.1898976 | 0.0885795 | 2.143809 | 0.0338270 | 0.0147262 | 0.3650690 |
| 0.1884328 | 0.0876465 | 2.149918 | 0.0333308 | 0.0151065 | 0.3617591 |
| 0.1899570 | 0.0841901 | 2.256288 | 0.0256482 | 0.0234661 | 0.3564480 |
| 0.1856386 | 0.0897414 | 2.068596 | 0.0404643 | 0.0081812 | 0.3630960 |
| 0.1897209 | 0.0881998 | 2.151035 | 0.0332278 | 0.0153118 | 0.3641299 |
| 0.1895477 | 0.0874451 | 2.167619 | 0.0319157 | 0.0166310 | 0.3624644 |
| 0.1896174 | 0.0883108 | 2.147161 | 0.0335540 | 0.0149775 | 0.3642573 |
| 0.1895146 | 0.0885351 | 2.140560 | 0.0340804 | 0.0144426 | 0.3645866 |
| 0.1895751 | 0.0878098 | 2.158929 | 0.0326105 | 0.0159259 | 0.3632243 |
| 0.1917748 | 0.0855142 | 2.242608 | 0.0265285 | 0.0226763 | 0.3608732 |
| 0.1880524 | 0.0919729 | 2.044649 | 0.0428044 | 0.0061822 | 0.3699225 |
| 0.1891769 | 0.0872782 | 2.167515 | 0.0319496 | 0.0165674 | 0.3617863 |
| 0.1908020 | 0.0914758 | 2.085819 | 0.0388634 | 0.0099030 | 0.3717009 |
| 0.1894118 | 0.0888204 | 2.132526 | 0.0347470 | 0.0137756 | 0.3650481 |
| 0.1994114 | 0.0907991 | 2.196184 | 0.0297610 | 0.0198625 | 0.3789603 |
| 0.1899018 | 0.0876054 | 2.167696 | 0.0319098 | 0.0166682 | 0.3631354 |
| 0.1888038 | 0.0894890 | 2.109800 | 0.0367075 | 0.0118339 | 0.3657737 |
| 0.1925240 | 0.0852852 | 2.257412 | 0.0255879 | 0.0238560 | 0.3611919 |
| 0.1899587 | 0.0877059 | 2.165860 | 0.0320528 | 0.0165263 | 0.3633911 |
| 0.1897574 | 0.0880179 | 2.155894 | 0.0328516 | 0.0156966 | 0.3638182 |
| 0.1900257 | 0.0874075 | 2.174021 | 0.0314341 | 0.0171721 | 0.3628792 |
| 0.1900081 | 0.0882910 | 2.152067 | 0.0331578 | 0.0154073 | 0.3646090 |
| 0.1900366 | 0.0883163 | 2.151771 | 0.0331816 | 0.0153856 | 0.3646875 |
| 0.1870585 | 0.0908157 | 2.059760 | 0.0413146 | 0.0074768 | 0.3666403 |
| 0.1913818 | 0.0880400 | 2.173804 | 0.0314381 | 0.0172886 | 0.3654749 |
| 0.1920468 | 0.0860894 | 2.230782 | 0.0273707 | 0.0217652 | 0.3623284 |
| 0.1927610 | 0.0853660 | 2.258053 | 0.0255233 | 0.0239556 | 0.3615664 |
| 0.1892257 | 0.0866435 | 2.183956 | 0.0306674 | 0.0178941 | 0.3605573 |
| 0.1893363 | 0.0856187 | 2.211390 | 0.0286787 | 0.0200201 | 0.3586525 |
| 0.1872263 | 0.0872593 | 2.145631 | 0.0336784 | 0.0146657 | 0.3597869 |
| 0.1863281 | 0.0921198 | 2.022672 | 0.0450530 | 0.0041676 | 0.3684886 |
| 0.1902387 | 0.0875720 | 2.172370 | 0.0315483 | 0.0170712 | 0.3634063 |
| 0.1905243 | 0.0871710 | 2.185640 | 0.0305412 | 0.0181497 | 0.3628988 |
| 0.1890701 | 0.0906983 | 2.084605 | 0.0389754 | 0.0097087 | 0.3684314 |
| 0.1892476 | 0.0887162 | 2.133180 | 0.0346923 | 0.0138174 | 0.3646778 |
| 0.1934850 | 0.0884912 | 2.186488 | 0.0304904 | 0.0184882 | 0.3684818 |
| 0.1968182 | 0.1151566 | 1.709135 | 0.0897237 | -0.0309261 | 0.4245626 |
| 0.1890500 | 0.0863240 | 2.190005 | 0.0302286 | 0.0183391 | 0.3597610 |
| 0.1885415 | 0.0894374 | 2.108085 | 0.0368582 | 0.0116737 | 0.3654093 |
| 0.1892877 | 0.0855761 | 2.211923 | 0.0286410 | 0.0200558 | 0.3585197 |
| 0.1896529 | 0.0889030 | 2.133256 | 0.0346860 | 0.0138533 | 0.3654525 |
| 0.1894624 | 0.0882922 | 2.145858 | 0.0336599 | 0.0148593 | 0.3640655 |
| 0.1917401 | 0.0935210 | 2.050237 | 0.0422624 | 0.0067967 | 0.3766835 |
| 0.1894905 | 0.0885835 | 2.139117 | 0.0342124 | 0.0143112 | 0.3646698 |
| 0.1894695 | 0.0876072 | 2.162716 | 0.0323118 | 0.0162209 | 0.3627180 |
| 0.1897645 | 0.0880294 | 2.155695 | 0.0328546 | 0.0156924 | 0.3638366 |
| 0.1898430 | 0.0861527 | 2.203564 | 0.0292376 | 0.0194708 | 0.3602151 |
| 0.1935889 | 0.0897753 | 2.156370 | 0.0328136 | 0.0160527 | 0.3711251 |
| 0.1901114 | 0.0835333 | 2.275874 | 0.0244164 | 0.0249191 | 0.3553036 |
| 0.1921936 | 0.0928420 | 2.070115 | 0.0403335 | 0.0085929 | 0.3757942 |
| 0.1897451 | 0.0880317 | 2.155418 | 0.0328895 | 0.0156571 | 0.3638332 |
| 0.1892785 | 0.0882959 | 2.143685 | 0.0338372 | 0.0146681 | 0.3638889 |
| 0.1902220 | 0.0870065 | 2.186297 | 0.0304921 | 0.0181727 | 0.3622713 |
| 0.1886690 | 0.0866849 | 2.176493 | 0.0312451 | 0.0172444 | 0.3600937 |
| 0.1905635 | 0.0856221 | 2.225634 | 0.0276974 | 0.0212293 | 0.3598978 |
| 0.1895705 | 0.0884484 | 2.143288 | 0.0338565 | 0.0146698 | 0.3644712 |
| 0.1914289 | 0.0855009 | 2.238912 | 0.0267741 | 0.0223569 | 0.3605010 |
| 0.1898751 | 0.0880624 | 2.156143 | 0.0328188 | 0.0157378 | 0.3640125 |
| 0.1898301 | 0.0880984 | 2.154751 | 0.0329299 | 0.0156216 | 0.3640386 |
| 0.1899318 | 0.0880288 | 2.157608 | 0.0327023 | 0.0158608 | 0.3640027 |
| 0.1905511 | 0.0902907 | 2.110419 | 0.0366532 | 0.0119959 | 0.3691064 |
| 0.1902141 | 0.0858846 | 2.214764 | 0.0284283 | 0.0203833 | 0.3600450 |
| 0.1891178 | 0.0888197 | 2.129233 | 0.0350501 | 0.0134598 | 0.3647758 |
| 0.1874340 | 0.0824692 | 2.272774 | 0.0246078 | 0.0243460 | 0.3505219 |
| 0.1892929 | 0.0877294 | 2.157690 | 0.0326958 | 0.0158140 | 0.3627719 |
| 0.1898535 | 0.0866550 | 2.190912 | 0.0301615 | 0.0184879 | 0.3612191 |
| 0.1897554 | 0.0881877 | 2.151723 | 0.0331725 | 0.0153704 | 0.3641405 |
| 0.1899841 | 0.0867373 | 2.190340 | 0.0301913 | 0.0184671 | 0.3615012 |
| 0.1896501 | 0.0879459 | 2.156441 | 0.0327951 | 0.0157432 | 0.3635570 |
| 0.1897206 | 0.0880195 | 2.155438 | 0.0328879 | 0.0156567 | 0.3637846 |
| 0.1892194 | 0.0863928 | 2.190221 | 0.0302001 | 0.0183836 | 0.3600553 |
| 0.1893224 | 0.0875846 | 2.161595 | 0.0323871 | 0.0161299 | 0.3625149 |
| 0.1898400 | 0.0880539 | 2.155954 | 0.0328339 | 0.0157196 | 0.3639605 |
| 0.1794748 | 0.0857652 | 2.092630 | 0.0382404 | 0.0098689 | 0.3490807 |
| 0.1900616 | 0.0846729 | 2.244658 | 0.0264050 | 0.0226159 | 0.3575074 |
| 0.1903354 | 0.0871325 | 2.184437 | 0.0306313 | 0.0180369 | 0.3626339 |
| 0.1898280 | 0.0886914 | 2.140320 | 0.0341132 | 0.0144354 | 0.3652207 |
| 0.1902226 | 0.0865903 | 2.196813 | 0.0297275 | 0.0189851 | 0.3614602 |
| 0.1903063 | 0.0893678 | 2.129474 | 0.0350164 | 0.0135761 | 0.3670365 |
| 0.1905583 | 0.0906505 | 2.102120 | 0.0373729 | 0.0113031 | 0.3698135 |
| 0.1896650 | 0.0877076 | 2.162469 | 0.0323442 | 0.0162064 | 0.3631237 |
| 0.1900185 | 0.0885966 | 2.144761 | 0.0337362 | 0.0148248 | 0.3652122 |
| 0.1895990 | 0.0886836 | 2.137927 | 0.0342976 | 0.0142334 | 0.3649647 |
| 0.1892893 | 0.0880145 | 2.150660 | 0.0332710 | 0.0152352 | 0.3633433 |
| 0.1892568 | 0.0901717 | 2.098850 | 0.0376788 | 0.0109368 | 0.3675768 |
| 0.1899536 | 0.0894095 | 2.124534 | 0.0354346 | 0.0131408 | 0.3667664 |
| 0.1896331 | 0.0882847 | 2.147972 | 0.0334752 | 0.0150562 | 0.3642099 |
| 0.1897318 | 0.0880843 | 2.153979 | 0.0329915 | 0.0155511 | 0.3639125 |
| 0.1899094 | 0.0892838 | 2.127031 | 0.0352227 | 0.0133453 | 0.3664736 |
| 0.1896161 | 0.0887412 | 2.136731 | 0.0344098 | 0.0141250 | 0.3651072 |
| 0.1893939 | 0.0874822 | 2.164942 | 0.0321245 | 0.0164039 | 0.3623840 |
| 0.1893680 | 0.0857105 | 2.209391 | 0.0288206 | 0.0198703 | 0.3588657 |
| 0.1895283 | 0.0863523 | 2.194826 | 0.0298730 | 0.0187613 | 0.3602952 |
| 0.1896672 | 0.0883806 | 2.146028 | 0.0336460 | 0.0148892 | 0.3644451 |
| 0.1897249 | 0.0882227 | 2.150522 | 0.0332691 | 0.0152705 | 0.3641792 |
| 0.1906219 | 0.0853974 | 2.232174 | 0.0272391 | 0.0217433 | 0.3595005 |
| 0.1896140 | 0.0884483 | 2.143784 | 0.0338290 | 0.0147022 | 0.3645258 |
| 0.1912048 | 0.0844211 | 2.264893 | 0.0251004 | 0.0242569 | 0.3581527 |
| 0.1895172 | 0.0886370 | 2.138126 | 0.0342943 | 0.0142320 | 0.3648023 |
| 0.1899587 | 0.0874207 | 2.172925 | 0.0315183 | 0.0170789 | 0.3628386 |
| 0.1906898 | 0.0884795 | 2.155188 | 0.0329079 | 0.0157163 | 0.3656633 |
| 0.1885328 | 0.0871918 | 2.162278 | 0.0323462 | 0.0161058 | 0.3609598 |
| 0.1894506 | 0.0882631 | 2.146431 | 0.0336133 | 0.0149050 | 0.3639963 |
| 0.1900917 | 0.0884407 | 2.149370 | 0.0333751 | 0.0151949 | 0.3649886 |
| 0.1874470 | 0.0850272 | 2.204553 | 0.0291665 | 0.0193006 | 0.3555934 |
| 0.1911127 | 0.0882898 | 2.164607 | 0.0321635 | 0.0165143 | 0.3657111 |
| 0.1895040 | 0.0884554 | 2.142368 | 0.0339319 | 0.0145895 | 0.3644185 |
| 0.1907109 | 0.0854679 | 2.231373 | 0.0273056 | 0.0216816 | 0.3597402 |
| 0.1893340 | 0.0917568 | 2.063432 | 0.0409594 | 0.0078912 | 0.3707768 |
| 0.1898826 | 0.0877062 | 2.164986 | 0.0321211 | 0.0164497 | 0.3633156 |
| 0.1897615 | 0.0877221 | 2.163213 | 0.0322599 | 0.0162971 | 0.3632258 |
| 0.1896416 | 0.0884668 | 2.143646 | 0.0338272 | 0.0147045 | 0.3645787 |
| 0.1897274 | 0.0882802 | 2.149151 | 0.0333798 | 0.0151594 | 0.3642955 |
| 0.1874839 | 0.0871073 | 2.152334 | 0.0331364 | 0.0152240 | 0.3597438 |
| 0.1896744 | 0.0876632 | 2.163673 | 0.0322367 | 0.0163151 | 0.3630336 |
| 0.1894249 | 0.0881903 | 2.147911 | 0.0334932 | 0.0150232 | 0.3638266 |
| 0.1896158 | 0.0879146 | 2.156818 | 0.0327780 | 0.0157594 | 0.3634722 |
| 0.1886234 | 0.0835740 | 2.256963 | 0.0256048 | 0.0233507 | 0.3538960 |
| 0.1892933 | 0.0897818 | 2.108370 | 0.0368331 | 0.0117443 | 0.3668423 |
| 0.1885648 | 0.0831488 | 2.267800 | 0.0249177 | 0.0241331 | 0.3529965 |
| 0.1893586 | 0.0893978 | 2.118157 | 0.0359808 | 0.0125690 | 0.3661482 |
| 0.1893304 | 0.0890414 | 2.126318 | 0.0352831 | 0.0132455 | 0.3654153 |
| 0.1894142 | 0.0889088 | 2.130432 | 0.0349358 | 0.0135916 | 0.3652368 |
| 0.1890185 | 0.0880312 | 2.147177 | 0.0335527 | 0.0149315 | 0.3631055 |
| 0.1852496 | 0.0794997 | 2.330194 | 0.0212675 | 0.0280342 | 0.3424651 |
| 0.1886422 | 0.0860301 | 2.192747 | 0.0300134 | 0.0185236 | 0.3587608 |
| 0.1896607 | 0.0875926 | 2.165258 | 0.0321126 | 0.0164409 | 0.3628804 |
| 0.1890887 | 0.0881816 | 2.144309 | 0.0337731 | 0.0147156 | 0.3634618 |
| 0.1900426 | 0.0910718 | 2.086734 | 0.0387792 | 0.0099426 | 0.3701427 |
| 0.1897864 | 0.0879173 | 2.158691 | 0.0326164 | 0.0159359 | 0.3636369 |
| 0.1876850 | 0.0879293 | 2.134500 | 0.0345953 | 0.0137995 | 0.3615704 |
| 0.1898034 | 0.0880563 | 2.155477 | 0.0328848 | 0.0156666 | 0.3639401 |
| 0.1897518 | 0.0880519 | 2.154999 | 0.0329100 | 0.0156352 | 0.3638685 |
| 0.1890736 | 0.0848397 | 2.228598 | 0.0274822 | 0.0212979 | 0.3568494 |
| 0.1902266 | 0.0869692 | 2.187286 | 0.0304182 | 0.0182509 | 0.3622023 |
| 0.1901475 | 0.0864046 | 2.200663 | 0.0294473 | 0.0192771 | 0.3610179 |
| 0.1912187 | 0.0876875 | 2.180685 | 0.0309927 | 0.0177520 | 0.3646855 |
| 0.1897455 | 0.0867766 | 2.186599 | 0.0304695 | 0.0181509 | 0.3613402 |
| 0.1899367 | 0.0871532 | 2.179343 | 0.0310156 | 0.0175972 | 0.3622762 |
| 0.1895812 | 0.0876759 | 2.162295 | 0.0323448 | 0.0161968 | 0.3629656 |
ggplot(results.Leave1studyout)+stat_dots(aes(x=Estimate), alpha=0.8, dotsize=1.5, shape=16, show.legend=F, col="darkcyan") # Distribution of overall mean after removal of one study at a time
results.Leave1studyout %>%
summarise(estimate= mean(Estimate), # calculate the mean coefficients across the models generated
se=mean(SE),
zval=mean(zval),
pval=mean(pval),
ci.lb=mean(ci.lb),
ci.ub=mean(ci.ub)) %>%
kable()%>% kable_styling("striped", position="left")| estimate | se | zval | pval | ci.lb | ci.ub |
|---|---|---|---|---|---|
| 0.1897728 | 0.0880607 | 2.156954 | 0.0331004 | 0.0156305 | 0.363915 |
Leave one species out
d$species_ID<-as.factor(d$species_ID)
d<-as.data.frame(d) # Only work with a dataframe
VCV_matrix<-list() # will need new VCV matrices because the sample size will be iteratively reduced
Leave1speciesout<-list() # create a list that will host the results of each model
for(i in 1:length(levels(d$species_ID))){ # N models = N species
VCV_matrix[[i]]<-make_VCV_matrix(d[d$species_ID != levels(d$species_ID)[i], ], V="Var_dARR", cluster="shared_trt_ID", obs="es_ID") # Create a new VCV matrix for each new model
Leave1speciesout[[i]] <- rma.mv(yi = dARR, V = VCV_matrix[[i]], # Same model structure as all the models we fitted
method="REML",
test="t",
dfs="contain",
random = list(~1|species_ID,
~1|phylogeny,
~1|es_ID),
R= list(phylogeny = phylo_matrix),
data = d[d$species_ID != levels(d$species_ID)[i], ],
sparse=TRUE) # Generate a new model for each new data (iterative removal of one species at a time)
}
# The output is a list so we need to summarise the coefficients of all the models performed
results.Leave1speciesout<-as.data.frame(cbind(
sapply(Leave1speciesout, function(x) summary(x)$beta), # extract the beta coefficient from all models
sapply(Leave1speciesout, function(x) summary(x)$se), # extract the standard error from all models
sapply(Leave1speciesout, function(x) summary(x)$zval), # extract the z value from all models
sapply(Leave1speciesout, function(x) summary(x)$pval), # extract the p value from all models
sapply(Leave1speciesout, function(x) summary(x)$ci.lb), # extract the lower confidence interval for all models
sapply(Leave1speciesout, function(x) summary(x)$ci.ub))) # extract the upper confidence interval for all models
colnames(results.Leave1speciesout)=c("Estimate", "SE", "zval", "pval", "ci.lb", "ci.ub") # change column names
kable(results.Leave1speciesout)%>% kable_styling("striped", position="left") %>% scroll_box(width="100%", height="500px") # Table of the results from all models
ggplot(results.Leave1speciesout)+stat_dots(aes(x=Estimate), alpha=0.8, dotsize=1.5, shape=16, show.legend=F, col="darkcyan") # Distribution of overall mean after removal of one species at a time
results.Leave1speciesout %>%
summarise(estimate= mean(Estimate), # calculate the mean coefficients across the models generated
se=mean(SE),
zval=mean(zval),
pval=mean(pval),
ci.lb=mean(ci.lb),
ci.ub=mean(ci.ub)) %>%
kable()%>% kable_styling("striped", position="left")load(here("Rdata", "Leaveonespeciesout.RData")) # Load the output of the models to avoid running them
kable(results.Leave1speciesout)%>% kable_styling("striped", position="left") %>% scroll_box(width="100%", height="500px") # Table of the results from all models| Estimate | SE | zval | pval | ci.lb | ci.ub |
|---|---|---|---|---|---|
| 0.1924767 | 0.0834555 | 2.306339 | 0.0226034 | 0.0274384 | 0.3575151 |
| 0.1897048 | 0.0882498 | 2.149635 | 0.0333537 | 0.0151855 | 0.3642240 |
| 0.1908438 | 0.0901151 | 2.117779 | 0.0360134 | 0.0126357 | 0.3690520 |
| 0.1902537 | 0.0888599 | 2.141053 | 0.0340530 | 0.0145279 | 0.3659795 |
| 0.1898509 | 0.0882153 | 2.152133 | 0.0331525 | 0.0153999 | 0.3643019 |
| 0.1898976 | 0.0885795 | 2.143809 | 0.0338270 | 0.0147262 | 0.3650690 |
| 0.1884328 | 0.0876465 | 2.149918 | 0.0333308 | 0.0151065 | 0.3617591 |
| 0.1867680 | 0.0898680 | 2.078248 | 0.0395663 | 0.0090486 | 0.3644874 |
| 0.1896174 | 0.0883108 | 2.147161 | 0.0335540 | 0.0149775 | 0.3642573 |
| 0.1895751 | 0.0878098 | 2.158929 | 0.0326105 | 0.0159259 | 0.3632243 |
| 0.1867280 | 0.0941494 | 1.983316 | 0.0493463 | 0.0005419 | 0.3729141 |
| 0.1893618 | 0.0877713 | 2.157446 | 0.0327281 | 0.0157887 | 0.3629349 |
| 0.1894765 | 0.0872333 | 2.172067 | 0.0315843 | 0.0169675 | 0.3619856 |
| 0.1895850 | 0.0876949 | 2.161870 | 0.0323783 | 0.0161630 | 0.3630071 |
| 0.1908020 | 0.0914758 | 2.085819 | 0.0388634 | 0.0099030 | 0.3717009 |
| 0.1904728 | 0.0874760 | 2.177428 | 0.0311738 | 0.0174836 | 0.3634620 |
| 0.1905516 | 0.0875934 | 2.175410 | 0.0313278 | 0.0173303 | 0.3637729 |
| 0.1897574 | 0.0880179 | 2.155894 | 0.0328516 | 0.0156966 | 0.3638182 |
| 0.1900257 | 0.0874075 | 2.174021 | 0.0314341 | 0.0171721 | 0.3628792 |
| 0.1900081 | 0.0882910 | 2.152067 | 0.0331578 | 0.0154073 | 0.3646090 |
| 0.1900366 | 0.0883163 | 2.151771 | 0.0331816 | 0.0153856 | 0.3646875 |
| 0.1920676 | 0.0857235 | 2.240549 | 0.0266771 | 0.0225442 | 0.3615910 |
| 0.1893830 | 0.0885411 | 2.138929 | 0.0342279 | 0.0142877 | 0.3644784 |
| 0.1899570 | 0.0841901 | 2.256288 | 0.0256482 | 0.0234661 | 0.3564480 |
| 0.1903609 | 0.0876692 | 2.171354 | 0.0316393 | 0.0169897 | 0.3637320 |
| 0.1901264 | 0.0875716 | 2.171097 | 0.0316591 | 0.0169483 | 0.3633045 |
| 0.1900914 | 0.0878720 | 2.163276 | 0.0322678 | 0.0163192 | 0.3638636 |
| 0.1900200 | 0.0880308 | 2.158563 | 0.0326395 | 0.0159337 | 0.3641063 |
| 0.1893363 | 0.0856187 | 2.211390 | 0.0286787 | 0.0200201 | 0.3586525 |
| 0.1872263 | 0.0872593 | 2.145631 | 0.0336784 | 0.0146657 | 0.3597869 |
| 0.1846879 | 0.0928533 | 1.989030 | 0.0487044 | 0.0010650 | 0.3683109 |
| 0.1890701 | 0.0906983 | 2.084605 | 0.0389754 | 0.0097087 | 0.3684314 |
| 0.1934850 | 0.0884912 | 2.186488 | 0.0304904 | 0.0184882 | 0.3684818 |
| 0.1924329 | 0.0985551 | 1.952541 | 0.0529279 | -0.0024658 | 0.3873316 |
| 0.1931318 | 0.0998147 | 1.934903 | 0.0550780 | -0.0042578 | 0.3905215 |
| 0.1890500 | 0.0863240 | 2.190005 | 0.0302286 | 0.0183391 | 0.3597610 |
| 0.1885415 | 0.0894374 | 2.108085 | 0.0368582 | 0.0116737 | 0.3654093 |
| 0.1892877 | 0.0855761 | 2.211923 | 0.0286410 | 0.0200558 | 0.3585197 |
| 0.1894624 | 0.0882922 | 2.145858 | 0.0336599 | 0.0148593 | 0.3640655 |
| 0.1917401 | 0.0935210 | 2.050237 | 0.0422624 | 0.0067967 | 0.3766835 |
| 0.1895160 | 0.0885735 | 2.139646 | 0.0341688 | 0.0143565 | 0.3646755 |
| 0.1893531 | 0.0884143 | 2.141658 | 0.0340032 | 0.0145085 | 0.3641977 |
| 0.1898430 | 0.0861527 | 2.203564 | 0.0292376 | 0.0194708 | 0.3602151 |
| 0.1935889 | 0.0897753 | 2.156370 | 0.0328136 | 0.0160527 | 0.3711251 |
| 0.1901114 | 0.0835333 | 2.275874 | 0.0244164 | 0.0249191 | 0.3553036 |
| 0.1903242 | 0.0873715 | 2.178334 | 0.0311048 | 0.0175419 | 0.3631066 |
| 0.1893566 | 0.0887014 | 2.134765 | 0.0345733 | 0.0139442 | 0.3647690 |
| 0.1899446 | 0.0888877 | 2.136906 | 0.0343953 | 0.0141638 | 0.3657254 |
| 0.1916686 | 0.0859927 | 2.228893 | 0.0274620 | 0.0216128 | 0.3617245 |
| 0.1895825 | 0.0875805 | 2.164667 | 0.0321588 | 0.0163868 | 0.3627782 |
| 0.1921936 | 0.0928420 | 2.070115 | 0.0403335 | 0.0085929 | 0.3757942 |
| 0.1897451 | 0.0880317 | 2.155418 | 0.0328895 | 0.0156571 | 0.3638332 |
| 0.1881906 | 0.0825785 | 2.278930 | 0.0242290 | 0.0248866 | 0.3514946 |
| 0.1886690 | 0.0866849 | 2.176493 | 0.0312451 | 0.0172444 | 0.3600937 |
| 0.1892308 | 0.0880637 | 2.148794 | 0.0334217 | 0.0150794 | 0.3633821 |
| 0.1897710 | 0.0878699 | 2.159681 | 0.0325510 | 0.0160029 | 0.3635390 |
| 0.1894221 | 0.0886188 | 2.137493 | 0.0343467 | 0.0141730 | 0.3646713 |
| 0.1898464 | 0.0872887 | 2.174926 | 0.0313648 | 0.0172278 | 0.3624651 |
| 0.1905001 | 0.0863099 | 2.207165 | 0.0289793 | 0.0198171 | 0.3611831 |
| 0.1905511 | 0.0902907 | 2.110419 | 0.0366532 | 0.0119959 | 0.3691064 |
| 0.1899535 | 0.0864437 | 2.197423 | 0.0296829 | 0.0190057 | 0.3609012 |
| 0.1891397 | 0.0888859 | 2.127893 | 0.0351498 | 0.0133624 | 0.3649170 |
| 0.1897736 | 0.0880071 | 2.156343 | 0.0328158 | 0.0157341 | 0.3638131 |
| 0.1874340 | 0.0824692 | 2.272774 | 0.0246078 | 0.0243460 | 0.3505219 |
| 0.1908732 | 0.0858545 | 2.223216 | 0.0278516 | 0.0210906 | 0.3606557 |
| 0.1898535 | 0.0866550 | 2.190912 | 0.0301615 | 0.0184879 | 0.3612191 |
| 0.1897371 | 0.0881657 | 2.152052 | 0.0331590 | 0.0153842 | 0.3640900 |
| 0.1899333 | 0.0869840 | 2.183544 | 0.0307110 | 0.0179173 | 0.3619494 |
| 0.1897206 | 0.0880195 | 2.155438 | 0.0328879 | 0.0156567 | 0.3637846 |
| 0.1892903 | 0.0871312 | 2.172476 | 0.0315529 | 0.0169832 | 0.3615975 |
| 0.1794748 | 0.0857652 | 2.092630 | 0.0382404 | 0.0098689 | 0.3490807 |
| 0.1900616 | 0.0846729 | 2.244658 | 0.0264050 | 0.0226159 | 0.3575074 |
| 0.1898280 | 0.0886914 | 2.140320 | 0.0341132 | 0.0144354 | 0.3652207 |
| 0.1902226 | 0.0865903 | 2.196813 | 0.0297275 | 0.0189851 | 0.3614602 |
| 0.1903063 | 0.0893678 | 2.129474 | 0.0350164 | 0.0135761 | 0.3670365 |
| 0.1904055 | 0.0862630 | 2.207267 | 0.0289720 | 0.0198151 | 0.3609959 |
| 0.1911552 | 0.0925271 | 2.065937 | 0.0407325 | 0.0081772 | 0.3741332 |
| 0.1904730 | 0.0844938 | 2.254283 | 0.0257773 | 0.0233813 | 0.3575646 |
| 0.1897109 | 0.0881263 | 2.152716 | 0.0331057 | 0.0154358 | 0.3639861 |
| 0.1896835 | 0.0882637 | 2.149055 | 0.0334005 | 0.0151367 | 0.3642303 |
| 0.1892072 | 0.0908954 | 2.081593 | 0.0392544 | 0.0094561 | 0.3689583 |
| 0.1892893 | 0.0880145 | 2.150660 | 0.0332710 | 0.0152352 | 0.3633433 |
| 0.1892568 | 0.0901717 | 2.098850 | 0.0376788 | 0.0109368 | 0.3675768 |
| 0.1899536 | 0.0894095 | 2.124534 | 0.0354346 | 0.0131408 | 0.3667664 |
| 0.1899094 | 0.0892838 | 2.127031 | 0.0352227 | 0.0133453 | 0.3664736 |
| 0.1896161 | 0.0887412 | 2.136731 | 0.0344098 | 0.0141250 | 0.3651072 |
| 0.1902362 | 0.0867392 | 2.193198 | 0.0299927 | 0.0187042 | 0.3617683 |
| 0.1860602 | 0.0799310 | 2.327760 | 0.0214005 | 0.0279918 | 0.3441287 |
| 0.1893680 | 0.0857105 | 2.209391 | 0.0288206 | 0.0198703 | 0.3588657 |
| 0.1895283 | 0.0863523 | 2.194826 | 0.0298730 | 0.0187613 | 0.3602952 |
| 0.1896672 | 0.0883806 | 2.146028 | 0.0336460 | 0.0148892 | 0.3644451 |
| 0.1906219 | 0.0853974 | 2.232174 | 0.0272391 | 0.0217433 | 0.3595005 |
| 0.1896140 | 0.0884483 | 2.143784 | 0.0338290 | 0.0147022 | 0.3645258 |
| 0.1895172 | 0.0886370 | 2.138126 | 0.0342943 | 0.0142320 | 0.3648023 |
| 0.1899587 | 0.0874207 | 2.172925 | 0.0315183 | 0.0170789 | 0.3628386 |
| 0.1906898 | 0.0884795 | 2.155188 | 0.0329079 | 0.0157163 | 0.3656633 |
| 0.1885328 | 0.0871918 | 2.162278 | 0.0323462 | 0.0161058 | 0.3609598 |
| 0.1912048 | 0.0844211 | 2.264893 | 0.0251004 | 0.0242569 | 0.3581527 |
| 0.1894506 | 0.0882631 | 2.146431 | 0.0336133 | 0.0149050 | 0.3639963 |
| 0.1900917 | 0.0884407 | 2.149370 | 0.0333751 | 0.0151949 | 0.3649886 |
| 0.1874470 | 0.0850272 | 2.204553 | 0.0291665 | 0.0193006 | 0.3555934 |
| 0.1911127 | 0.0882898 | 2.164607 | 0.0321635 | 0.0165143 | 0.3657111 |
| 0.1901108 | 0.0869487 | 2.186470 | 0.0304917 | 0.0181644 | 0.3620572 |
| 0.1901068 | 0.0869605 | 2.186128 | 0.0305173 | 0.0181371 | 0.3620765 |
| 0.1901130 | 0.0868949 | 2.187850 | 0.0303888 | 0.0182730 | 0.3619531 |
| 0.1874839 | 0.0871073 | 2.152334 | 0.0331364 | 0.0152240 | 0.3597438 |
| 0.1896744 | 0.0876632 | 2.163673 | 0.0322367 | 0.0163151 | 0.3630336 |
| 0.1894249 | 0.0881903 | 2.147911 | 0.0334932 | 0.0150232 | 0.3638266 |
| 0.1893406 | 0.0916130 | 2.066745 | 0.0406551 | 0.0081704 | 0.3705108 |
| 0.1896158 | 0.0879146 | 2.156818 | 0.0327780 | 0.0157594 | 0.3634722 |
| 0.1886234 | 0.0835740 | 2.256963 | 0.0256048 | 0.0233507 | 0.3538960 |
| 0.1892933 | 0.0897818 | 2.108370 | 0.0368331 | 0.0117443 | 0.3668423 |
| 0.1893586 | 0.0893978 | 2.118157 | 0.0359808 | 0.0125690 | 0.3661482 |
| 0.1893304 | 0.0890414 | 2.126318 | 0.0352831 | 0.0132455 | 0.3654153 |
| 0.1894142 | 0.0889088 | 2.130432 | 0.0349358 | 0.0135916 | 0.3652368 |
| 0.1890185 | 0.0880312 | 2.147177 | 0.0335527 | 0.0149315 | 0.3631055 |
| 0.1852496 | 0.0794997 | 2.330194 | 0.0212675 | 0.0280342 | 0.3424651 |
| 0.1885648 | 0.0831488 | 2.267800 | 0.0249177 | 0.0241331 | 0.3529965 |
| 0.1896607 | 0.0875926 | 2.165258 | 0.0321126 | 0.0164409 | 0.3628804 |
| 0.1980547 | 0.0903518 | 2.192039 | 0.0300781 | 0.0193785 | 0.3767310 |
| 0.1900426 | 0.0910718 | 2.086734 | 0.0387792 | 0.0099426 | 0.3701427 |
| 0.1876850 | 0.0879293 | 2.134500 | 0.0345953 | 0.0137995 | 0.3615704 |
| 0.1898034 | 0.0880563 | 2.155477 | 0.0328848 | 0.0156666 | 0.3639401 |
| 0.1900343 | 0.0856026 | 2.219958 | 0.0280774 | 0.0207499 | 0.3593187 |
| 0.1890736 | 0.0848397 | 2.228598 | 0.0274822 | 0.0212979 | 0.3568494 |
| 0.1901475 | 0.0864046 | 2.200663 | 0.0294473 | 0.0192771 | 0.3610179 |
| 0.1905272 | 0.0877197 | 2.172001 | 0.0315895 | 0.0170562 | 0.3639982 |
| 0.1900132 | 0.0880952 | 2.156906 | 0.0327710 | 0.0157995 | 0.3642269 |
| 0.1897120 | 0.0879491 | 2.157066 | 0.0327583 | 0.0157873 | 0.3636367 |
| 0.1898449 | 0.0881269 | 2.154223 | 0.0329850 | 0.0155687 | 0.3641212 |
| 0.1901572 | 0.0884399 | 2.150128 | 0.0333139 | 0.0152618 | 0.3650525 |
| 0.1894199 | 0.0877824 | 2.157835 | 0.0326972 | 0.0158249 | 0.3630148 |
| 0.1895812 | 0.0876759 | 2.162295 | 0.0323448 | 0.0161968 | 0.3629656 |
| 0.1935150 | 0.0922495 | 2.097736 | 0.0377789 | 0.0110861 | 0.3759439 |
| 0.1901109 | 0.0868494 | 2.188972 | 0.0303053 | 0.0183609 | 0.3618609 |
| 0.1898975 | 0.0887416 | 2.139893 | 0.0341484 | 0.0144056 | 0.3653894 |
| 0.1901953 | 0.0921256 | 2.064522 | 0.0408684 | 0.0080113 | 0.3723792 |
| 0.1893608 | 0.0895556 | 2.114450 | 0.0363015 | 0.0122592 | 0.3664625 |
ggplot(results.Leave1speciesout)+stat_dots(aes(x=Estimate), alpha=0.8, dotsize=1.5, shape=16, show.legend=F, col="darkcyan") # Distribution of overall mean after removal of one species at a time
results.Leave1speciesout %>%
summarise(estimate= mean(Estimate), # calculate the mean coefficients across the models generated
se=mean(SE),
zval=mean(zval),
pval=mean(pval),
ci.lb=mean(ci.lb),
ci.ub=mean(ci.ub)) %>%
kable()%>% kable_styling("striped", position="left")| estimate | se | zval | pval | ci.lb | ci.ub |
|---|---|---|---|---|---|
| 0.1897393 | 0.0879348 | 2.159462 | 0.0328797 | 0.0158429 | 0.3636356 |
Analyses with direct designs only
Habitat
Individual coefficients
d.direct <- filter(d, brought_common_temp == "no")
mod.direct_habitat <- run.model(d.direct, ~habitat - 1)
summary(mod.direct_habitat)##
## Multivariate Meta-Analysis Model (k = 866; method: REML)
##
## logLik Deviance AIC BIC AICc
## -207.4173 414.8347 424.8347 448.6425 424.9046
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0057 0.0758 121 no species_ID no
## sigma^2.2 0.0074 0.0861 121 no phylogeny yes
## sigma^2.3 0.0693 0.2633 866 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 864) = 37465.3848, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 119) = 12.3613, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## habitataquatic 0.2334 0.0511 4.5674 119 <.0001 0.1322 0.3346 ***
## habitatterrestrial 0.0727 0.0767 0.9473 119 0.3454 -0.0793 0.2246
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.direct_habitat, mod = "habitat", data = d.direct, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Aquatic 0.23341363 0.13222224 0.3346050 -0.3441887 0.8110160
## 2 Terrestrial 0.07268808 -0.07925124 0.2246274 -0.5159292 0.6613054my.orchard(mod.direct_habitat, xlab = "dARR", mod = "habitat", alpha = 0.1, data = d.direct,
group = "species_ID", whisker = 0.07)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.direct_habitat)^2), x = fitted(mod.direct_habitat)) +
geom_point() + geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Individual coefficients (with heteroscedasticity)
VCV_direct <- make_VCV_matrix(d.direct, V = "Var_dARR", cluster = "shared_trt_ID",
obs = "es_ID")
mod.direct_habitat_het <- rma.mv(yi = dARR, V = VCV_direct, mods = ~habitat - 1,
method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~habitat | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d.direct, sparse = TRUE)
AICc(mod.direct_habitat)## [1] 424.9044AICc(mod.direct_habitat_het)## [1] 341.5476summary(mod.direct_habitat_het)##
## Multivariate Meta-Analysis Model (k = 866; method: REML)
##
## logLik Deviance AIC BIC AICc
## -164.7249 329.4498 341.4498 370.0192 341.5478
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0060 0.0777 121 no species_ID no
## sigma^2.2 0.0063 0.0796 121 no phylogeny yes
##
## outer factor: es_ID (nlvls = 866)
## inner factor: habitat (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0776 0.2785 781 no aquatic
## tau^2.2 0.0085 0.0924 85 no terrestrial
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 864) = 37465.3848, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 119) = 14.5396, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## habitataquatic 0.2341 0.0480 4.8775 119 <.0001 0.1391 0.3292 ***
## habitatterrestrial 0.0738 0.0645 1.1447 119 0.2546 -0.0539 0.2016
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.direct_habitat_het, mod = "habitat", data = d.direct, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Aquatic 0.23412219 0.13907723 0.3291672 -0.3672915 0.8355359
## 2 Terrestrial 0.07384853 -0.05389323 0.2015903 -0.2396942 0.3873912r2_ml(mod.direct_habitat_het)## R2_marginal R2_conditional
## 0.1553241 0.5672050my.orchard(mod.direct_habitat_het, mod = "habitat", xlab = "dARR", alpha = 0.1, data = d.direct,
group = "species_ID", whisker = 0.07)
Mean direct effects with habitat weighted equally
mod.adjusted_direct <- mod_results(model = mod.direct_habitat_het, data = d.direct,
mod = "1", weights = "equal", group = "species_ID")
mod.adjusted_direct$mod_table## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.1539854 0.05381858 0.2541521 -0.4257827 0.7337534my.orchard(mod.adjusted_direct, mod = "1", xlab = "dARR", alpha = 0.1, data = d.direct,
group = "species_ID", whisker = 0.04)
Taxonomic groups
Individual coefficients
mod.direct_taxa <- run.model(d.direct, ~taxonomic_group - 1)
summary(mod.direct_taxa)##
## Multivariate Meta-Analysis Model (k = 866; method: REML)
##
## logLik Deviance AIC BIC AICc
## -204.7335 409.4670 425.4670 463.5318 425.6361
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0051 0.0714 121 no species_ID no
## sigma^2.2 0.0139 0.1180 121 no phylogeny yes
## sigma^2.3 0.0692 0.2631 866 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 861) = 37180.6026, p-val < .0001
##
## Test of Moderators (coefficients 1:5):
## F(df1 = 5, df2 = 116) = 3.4192, p-val = 0.0064
##
## Model Results:
##
## estimate se tval df pval
## taxonomic_groupamphibian 0.2247 0.1215 1.8496 116 0.0669
## taxonomic_groupaquatic invertebrate 0.1947 0.0928 2.0995 116 0.0379
## taxonomic_groupfish 0.2843 0.0915 3.1073 116 0.0024
## taxonomic_groupreptile 0.1580 0.1462 1.0810 116 0.2819
## taxonomic_groupterrestrial invertebrate 0.0306 0.1089 0.2809 116 0.7793
## ci.lb ci.ub
## taxonomic_groupamphibian -0.0159 0.4652 .
## taxonomic_groupaquatic invertebrate 0.0110 0.3785 *
## taxonomic_groupfish 0.1031 0.4656 **
## taxonomic_groupreptile -0.1315 0.4476
## taxonomic_groupterrestrial invertebrate -0.1851 0.2462
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.direct_taxa, mod = "taxonomic_group", data = d.direct, group = "species_ID")## name estimate lowerCL upperCL lowerPR
## 1 Amphibian 0.22465082 -0.01590805 0.4652097 -0.4110431
## 2 Aquatic invertebrate 0.19474979 0.01102876 0.3784708 -0.4216848
## 3 Fish 0.28434087 0.10309755 0.4655842 -0.3313598
## 4 Reptile 0.15803905 -0.13152214 0.4476002 -0.4977688
## 5 Terrestrial invertebrate 0.03058815 -0.18506361 0.2462399 -0.5961047
## upperPR
## 1 0.8603448
## 2 0.8111844
## 3 0.9000416
## 4 0.8138469
## 5 0.6572810my.orchard(mod.direct_taxa, xlab = "dARR", mod = "taxonomic_group", data = d.direct,
alpha = 0.1, group = "species_ID", whisker = 0.09)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.direct_taxa)^2), x = fitted(mod.direct_taxa)) + geom_point() +
geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Individual coefficients (with heteroscedasticity)
mod.direct_taxa_het <- rma.mv(yi = dARR, V = VCV_direct, mods = ~taxonomic_group -
1, method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~taxonomic_group | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d.direct, sparse = TRUE)
AICc(mod.direct_taxa)## [1] 425.6351AICc(mod.direct_taxa_het)## [1] 162.6418summary(mod.direct_taxa_het)##
## Multivariate Meta-Analysis Model (k = 866; method: REML)
##
## logLik Deviance AIC BIC AICc
## -69.1380 138.2760 162.2760 219.3732 162.6440
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0051 0.0716 121 no species_ID no
## sigma^2.2 0.0210 0.1449 121 no phylogeny yes
##
## outer factor: es_ID (nlvls = 866)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.1977 0.4446 65 no amphibian
## tau^2.2 0.1463 0.3824 208 no aquatic invertebrate
## tau^2.3 0.0318 0.1783 496 no fish
## tau^2.4 0.0046 0.0678 12 no reptile
## tau^2.5 0.0085 0.0921 85 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 861) = 37180.6026, p-val < .0001
##
## Test of Moderators (coefficients 1:5):
## F(df1 = 5, df2 = 116) = 2.8829, p-val = 0.0172
##
## Model Results:
##
## estimate se tval df pval
## taxonomic_groupamphibian 0.2165 0.1534 1.4113 116 0.1608
## taxonomic_groupaquatic invertebrate 0.2024 0.1154 1.7535 116 0.0822
## taxonomic_groupfish 0.2959 0.1084 2.7285 116 0.0074
## taxonomic_groupreptile 0.1432 0.1495 0.9577 116 0.3402
## taxonomic_groupterrestrial invertebrate 0.0309 0.1253 0.2468 116 0.8055
## ci.lb ci.ub
## taxonomic_groupamphibian -0.0873 0.5203
## taxonomic_groupaquatic invertebrate -0.0262 0.4310 .
## taxonomic_groupfish 0.0811 0.5106 **
## taxonomic_groupreptile -0.1529 0.4393
## taxonomic_groupterrestrial invertebrate -0.2172 0.2790
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.direct_taxa_het, mod = "taxonomic_group", data = d.direct, group = "species_ID")## name estimate lowerCL upperCL lowerPR
## 1 Amphibian 0.21650061 -0.08734726 0.5203485 -0.7685470
## 2 Aquatic invertebrate 0.20238847 -0.02621472 0.4309917 -0.6511378
## 3 Fish 0.29585838 0.08109085 0.5106259 -0.2269040
## 4 Reptile 0.14317390 -0.15290973 0.4392575 -0.3131056
## 5 Terrestrial invertebrate 0.03091705 -0.21721312 0.2790472 -0.4133293
## upperPR
## 1 1.2015482
## 2 1.0559147
## 3 0.8186208
## 4 0.5994534
## 5 0.4751634r2_ml(mod.direct_taxa_het)## R2_marginal R2_conditional
## 0.2015604 0.3583558my.orchard(mod.direct_taxa_het, mod = "taxonomic_group", xlab = "dARR", alpha = 0.1,
data = d.direct, group = "species_ID", whisker = 0.09)
Type of metric
Individual coefficients
mod.direct_metric <- run.model(d.direct, ~metric - 1)
summary(mod.direct_metric)##
## Multivariate Meta-Analysis Model (k = 866; method: REML)
##
## logLik Deviance AIC BIC AICc
## -209.0317 418.0635 428.0635 451.8713 428.1334
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0057 0.0754 121 no species_ID no
## sigma^2.2 0.0159 0.1263 121 no phylogeny yes
## sigma^2.3 0.0690 0.2627 866 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 864) = 47790.0581, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 864) = 6.0643, p-val = 0.0024
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## metricCTmax 0.2269 0.0722 3.1450 864 0.0017 0.0853 0.3685 **
## metricLT50 0.1753 0.0753 2.3275 864 0.0202 0.0275 0.3232 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.direct_metric, mod = "metric", data = d.direct, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 CTmax 0.2269223 0.08530761 0.3685370 -0.3807252 0.8345698
## 2 LT50 0.1753155 0.02747981 0.3231512 -0.4338118 0.7844429my.orchard(mod.direct_metric, xlab = "dARR", mod = "metric", alpha = 0.1, data = d.direct,
group = "species_ID", whisker = 0.07)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.direct_metric)^2), x = fitted(mod.direct_metric)) +
geom_point() + geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Individual coefficients (with heteroscedasticity)
mod.direct_metric_het <- rma.mv(yi = dARR, V = VCV_direct, mods = ~metric - 1, method = "REML",
test = "t", dfs = "contain", random = list(~1 | species_ID, ~1 | phylogeny, ~metric |
es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix), data = d.direct,
sparse = TRUE)
AICc(mod.direct_metric)## [1] 428.1332AICc(mod.direct_metric_het)## [1] 414.7547summary(mod.direct_metric_het)##
## Multivariate Meta-Analysis Model (k = 866; method: REML)
##
## logLik Deviance AIC BIC AICc
## -201.3285 402.6569 414.6569 443.2264 414.7550
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0061 0.0783 121 no species_ID no
## sigma^2.2 0.0159 0.1260 121 no phylogeny yes
##
## outer factor: es_ID (nlvls = 866)
## inner factor: metric (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0763 0.2762 647 no CTmax
## tau^2.2 0.0414 0.2035 219 no LT50
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 864) = 47790.0581, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 864) = 6.3190, p-val = 0.0019
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## metricCTmax 0.2250 0.0720 3.1247 864 0.0018 0.0837 0.3664 **
## metricLT50 0.1718 0.0742 2.3156 864 0.0208 0.0262 0.3173 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.direct_metric_het, mod = "metric", data = d.direct, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 CTmax 0.2250164 0.08367535 0.3663575 -0.4062437 0.8562766
## 2 LT50 0.1717566 0.02617401 0.3173391 -0.3433756 0.6868888r2_ml(mod.direct_metric_het)## R2_marginal R2_conditional
## 0.02381791 0.29568891my.orchard(mod.direct_metric_het, mod = "metric", xlab = "dARR", alpha = 0.1, data = d.direct,
group = "species_ID", whisker = 0.07)
Heating rate
Run model and plot results
mod.direct_ramping <- run.model(d.direct, ~scale(ramping, center = T, scale = F))
summary(mod.direct_ramping)##
## Multivariate Meta-Analysis Model (k = 653; method: REML)
##
## logLik Deviance AIC BIC AICc
## -147.7601 295.5202 305.5202 327.9127 305.6132
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0009 0.0301 99 no species_ID no
## sigma^2.2 0.0308 0.1755 99 no phylogeny yes
## sigma^2.3 0.0747 0.2732 653 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 651) = 41157.4313, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 651) = 0.1759, p-val = 0.6750
##
## Model Results:
##
## estimate se tval df pval
## intrcpt 0.2337 0.0990 2.3596 97 0.0203
## scale(ramping, center = T, scale = F) 0.0137 0.0327 0.4195 651 0.6750
## ci.lb ci.ub
## intrcpt 0.0371 0.4303 *
## scale(ramping, center = T, scale = F) -0.0505 0.0779
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1d.direct_ramping <- filter(d, ramping != "NA" & brought_common_temp == "no")
plot_continuous(d.direct_ramping, mod.direct_ramping, d.direct_ramping$ramping, "Heating rate (degrees/min)")
Analyses with persistent designs only
Habitat
Individual coefficients
d.persistent <- filter(d, brought_common_temp == "yes") # Only keep persistent data
mod.persistent_habitat <- run.model(d.persistent, ~habitat - 1)
summary(mod.persistent_habitat)##
## Multivariate Meta-Analysis Model (k = 223; method: REML)
##
## logLik Deviance AIC BIC AICc
## 28.5430 -57.0860 -47.0860 -30.0952 -46.8069
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0000 0.0000 23 no species_ID no
## sigma^2.2 0.0311 0.1765 23 no phylogeny yes
## sigma^2.3 0.0214 0.1461 223 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 221) = 6128.2502, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 21) = 0.1327, p-val = 0.8765
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## habitataquatic 0.0533 0.1049 0.5078 21 0.6169 -0.1649 0.2715
## habitatterrestrial 0.0519 0.1141 0.4549 21 0.6539 -0.1854 0.2891
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persistent_habitat, mod = "habitat", data = d.persistent, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Aquatic 0.05329268 -0.1649379 0.2715232 -0.4708093 0.5773947
## 2 Terrestrial 0.05189110 -0.1853570 0.2891393 -0.4804105 0.5841927my.orchard(mod.persistent_habitat, xlab = "dARR", mod = "habitat", alpha = 0.1, data = d.persistent,
group = "species_ID", whisker = 0.07)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.persistent_habitat)^2), x = fitted(mod.persistent_habitat)) +
geom_point() + geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Individual coefficients (with heteroscedasticity)
VCV_persistent <- make_VCV_matrix(d.persistent, V = "Var_dARR", cluster = "shared_trt_ID",
obs = "es_ID") # Re calculate VCV matrix with the new data
mod.persistent_habitat_het <- rma.mv(yi = dARR, V = VCV_persistent, mods = ~habitat -
1, method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~habitat | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d.persistent, sparse = TRUE)
AICc(mod.persistent_habitat)## [1] -46.80952AICc(mod.persistent_habitat_het)## [1] -64.95965summary(mod.persistent_habitat_het)##
## Multivariate Meta-Analysis Model (k = 223; method: REML)
##
## logLik Deviance AIC BIC AICc
## 38.6743 -77.3485 -65.3485 -44.9596 -64.9560
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0003 0.0167 23 no species_ID no
## sigma^2.2 0.0272 0.1650 23 no phylogeny yes
##
## outer factor: es_ID (nlvls = 223)
## inner factor: habitat (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0309 0.1758 148 no aquatic
## tau^2.2 0.0079 0.0889 75 no terrestrial
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 221) = 6128.2502, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 21) = 0.1348, p-val = 0.8746
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## habitataquatic 0.0507 0.0998 0.5075 21 0.6171 -0.1569 0.2582
## habitatterrestrial 0.0492 0.1066 0.4616 21 0.6491 -0.1726 0.2710
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persistent_habitat_het, mod = "habitat", data = d.persistent, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Aquatic 0.05065728 -0.1569176 0.2582322 -0.4930975 0.5944120
## 2 Terrestrial 0.04922949 -0.1725603 0.2710193 -0.4005826 0.4990416r2_ml(mod.persistent_habitat_het)## R2_marginal R2_conditional
## 1.662009e-05 1.015787e-02my.orchard(mod.persistent_habitat_het, mod = "habitat", xlab = "dARR", alpha = 0.1,
data = d.persistent, group = "species_ID", whisker = 0.07)
Mean persistent effects with habitat weighted equally
mod.adjusted_persistent <- mod_results(model = mod.persistent_habitat_het, data = d.persistent,
mod = "1", weights = "equal", group = "species_ID")
mod.adjusted_persistent$mod_table## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.04994338 -0.1521334 0.2520201 -0.4599315 0.5598183my.orchard(mod.adjusted_persistent, mod = "1", xlab = "dARR", alpha = 0.1, data = d.persistent,
group = "species_ID", whisker = 0.04)
Taxonomic groups
Individual coefficients
mod.persistent_taxa <- run.model(d.persistent, ~taxonomic_group - 1)
summary(mod.persistent_taxa)##
## Multivariate Meta-Analysis Model (k = 223; method: REML)
##
## logLik Deviance AIC BIC AICc
## 30.2273 -60.4546 -44.4546 -17.3786 -43.7656
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0000 0.0000 23 no species_ID no
## sigma^2.2 0.0402 0.2006 23 no phylogeny yes
## sigma^2.3 0.0214 0.1463 223 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 218) = 5315.6964, p-val < .0001
##
## Test of Moderators (coefficients 1:5):
## F(df1 = 5, df2 = 18) = 0.2834, p-val = 0.9161
##
## Model Results:
##
## estimate se tval df pval
## taxonomic_groupamphibian 0.0091 0.2120 0.0427 18 0.9664
## taxonomic_groupaquatic invertebrate 0.1523 0.1518 1.0029 18 0.3292
## taxonomic_groupfish -0.0279 0.1608 -0.1735 18 0.8642
## taxonomic_groupreptile -0.0511 0.1999 -0.2554 18 0.8013
## taxonomic_groupterrestrial invertebrate 0.1190 0.1647 0.7223 18 0.4794
## ci.lb ci.ub
## taxonomic_groupamphibian -0.4363 0.4544
## taxonomic_groupaquatic invertebrate -0.1667 0.4713
## taxonomic_groupfish -0.3657 0.3099
## taxonomic_groupreptile -0.4710 0.3689
## taxonomic_groupterrestrial invertebrate -0.2270 0.4650
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persistent_taxa, mod = "taxonomic_group", data = d.persistent, group = "species_ID")## name estimate lowerCL upperCL lowerPR
## 1 Amphibian 0.009050625 -0.4363420 0.4544433 -0.6768539
## 2 Aquatic invertebrate 0.152282841 -0.1667180 0.4712837 -0.4591515
## 3 Fish -0.027894294 -0.3657346 0.3099460 -0.6493655
## 4 Reptile -0.051058808 -0.4710038 0.3688862 -0.7207186
## 5 Terrestrial invertebrate 0.118957656 -0.2270469 0.4649622 -0.5069893
## upperPR
## 1 0.6949552
## 2 0.7637172
## 3 0.5935769
## 4 0.6186010
## 5 0.7449046my.orchard(mod.persistent_taxa, xlab = "dARR", mod = "taxonomic_group", alpha = 0.1,
data = d.persistent, group = "species_ID", whisker = 0.09)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.persistent_taxa)^2), x = fitted(mod.persistent_taxa)) +
geom_point() + geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Individual coefficients (with heteroscedasticity)
mod.persistent_taxa_het <- rma.mv(yi = dARR, V = VCV_persistent, mods = ~taxonomic_group -
1, method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~taxonomic_group | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d.persistent, sparse = TRUE)
AICc(mod.persistent_taxa)## [1] -43.78171AICc(mod.persistent_taxa_het)## [1] -75.11488summary(mod.persistent_taxa_het)##
## Multivariate Meta-Analysis Model (k = 223; method: REML)
##
## logLik Deviance AIC BIC AICc
## 50.3003 -100.6006 -76.6006 -35.9866 -75.0786
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0005 0.0216 23 no species_ID no
## sigma^2.2 0.0331 0.1821 23 no phylogeny yes
##
## outer factor: es_ID (nlvls = 223)
## inner factor: taxonomic_group (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0626 0.2501 6 no amphibian
## tau^2.2 0.0081 0.0899 13 no aquatic invertebrate
## tau^2.3 0.0332 0.1822 127 no fish
## tau^2.4 0.0309 0.1758 15 no reptile
## tau^2.5 0.0046 0.0675 62 no terrestrial invertebrate
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 218) = 5315.6964, p-val < .0001
##
## Test of Moderators (coefficients 1:5):
## F(df1 = 5, df2 = 18) = 0.2830, p-val = 0.9163
##
## Model Results:
##
## estimate se tval df pval
## taxonomic_groupamphibian 0.0050 0.2135 0.0235 18 0.9815
## taxonomic_groupaquatic invertebrate 0.1358 0.1350 1.0063 18 0.3276
## taxonomic_groupfish -0.0294 0.1469 -0.2003 18 0.8435
## taxonomic_groupreptile -0.0474 0.1848 -0.2563 18 0.8006
## taxonomic_groupterrestrial invertebrate 0.0968 0.1461 0.6625 18 0.5161
## ci.lb ci.ub
## taxonomic_groupamphibian -0.4436 0.4536
## taxonomic_groupaquatic invertebrate -0.1478 0.4194
## taxonomic_groupfish -0.3380 0.2791
## taxonomic_groupreptile -0.4357 0.3409
## taxonomic_groupterrestrial invertebrate -0.2102 0.4038
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persistent_taxa_het, mod = "taxonomic_group", data = d.persistent,
group = "species_ID")## name estimate lowerCL upperCL lowerPR
## 1 Amphibian 0.005026263 -0.4435812 0.4536337 -0.7860333
## 2 Aquatic invertebrate 0.135833415 -0.1477534 0.4194202 -0.3784558
## 3 Fish -0.029423673 -0.3379827 0.2791354 -0.6540624
## 4 Reptile -0.047365567 -0.4356609 0.3409298 -0.7072814
## 5 Terrestrial invertebrate 0.096803200 -0.2101875 0.4037939 -0.4157655
## upperPR
## 1 0.7960858
## 2 0.6501226
## 3 0.5952151
## 4 0.6125503
## 5 0.6093719r2_ml(mod.persistent_taxa_het)## R2_marginal R2_conditional
## 0.1091420 0.1215625my.orchard(mod.persistent_taxa_het, mod = "taxonomic_group", xlab = "dARR", alpha = 0.1,
data = d.persistent, group = "species_ID", whisker = 0.09)
Type of metric
Individual coefficients
mod.persistent_metric <- run.model(d.persistent, ~metric - 1)
summary(mod.persistent_metric)##
## Multivariate Meta-Analysis Model (k = 223; method: REML)
##
## logLik Deviance AIC BIC AICc
## 28.4387 -56.8773 -46.8773 -29.8865 -46.5982
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0000 0.0000 23 no species_ID no
## sigma^2.2 0.0268 0.1636 23 no phylogeny yes
## sigma^2.3 0.0215 0.1466 223 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 221) = 8562.9957, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 221) = 0.5988, p-val = 0.5503
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## metricCTmax 0.0441 0.0958 0.4602 221 0.6458 -0.1447 0.2329
## metricLT50 0.1483 0.1396 1.0626 221 0.2891 -0.1267 0.4233
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persistent_metric, mod = "metric", data = d.persistent, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 CTmax 0.04408739 -0.1446967 0.2328715 -0.4282615 0.5164363
## 2 LT50 0.14829967 -0.1267440 0.4233433 -0.3646556 0.6612549my.orchard(mod.persistent_metric, xlab = "dARR", mod = "metric", alpha = 0.1, data = d.persistent,
group = "species_ID", whisker = 0.07)
Heteroscedasticity check
qplot(y = sqrt(residuals(mod.persistent_metric)^2), x = fitted(mod.persistent_metric)) +
geom_point() + geom_smooth(method = "lm") + geom_hline(yintercept = 0, colour = "red")
Individual coefficients (with heteroscedasticity)
mod.persistent_metric_het <- rma.mv(yi = dARR, V = VCV_persistent, mods = ~metric -
1, method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~metric | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d.persistent, sparse = TRUE)
AICc(mod.persistent_metric)## [1] -46.60081AICc(mod.persistent_metric_het)## [1] -44.8704summary(mod.persistent_metric_het)##
## Multivariate Meta-Analysis Model (k = 223; method: REML)
##
## logLik Deviance AIC BIC AICc
## 28.6296 -57.2593 -45.2593 -24.8703 -44.8668
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0000 0.0000 23 no species_ID no
## sigma^2.2 0.0281 0.1677 23 no phylogeny yes
##
## outer factor: es_ID (nlvls = 223)
## inner factor: metric (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0213 0.1460 216 no CTmax
## tau^2.2 0.0524 0.2289 7 no LT50
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 221) = 8562.9957, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 221) = 0.6520, p-val = 0.5220
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## metricCTmax 0.0486 0.0982 0.4954 221 0.6208 -0.1449 0.2422
## metricLT50 0.1877 0.1658 1.1321 221 0.2588 -0.1390 0.5144
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.persistent_metric_het, mod = "metric", data = d.persistent, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 CTmax 0.04864284 -0.1448687 0.2421544 -0.4303011 0.5275867
## 2 LT50 0.18768761 -0.1390454 0.5144206 -0.4599910 0.8353662r2_ml(mod.persistent_metric_het)## R2_marginal R2_conditional
## 0.02056887 0.02056887my.orchard(mod.persistent_metric_het, mod = "metric", xlab = "dARR", alpha = 0.1,
data = d.persistent, group = "species_ID", whisker = 0.07)
Heating rate
Run model and plot results
mod.persistent_ramping <- run.model(filter(d, brought_common_temp == "yes"), ~scale(ramping,
center = T, scale = F))
summary(mod.persistent_ramping)##
## Multivariate Meta-Analysis Model (k = 202; method: REML)
##
## logLik Deviance AIC BIC AICc
## 27.8184 -55.6369 -45.6369 -29.1453 -45.3276
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0018 0.0422 20 no species_ID no
## sigma^2.2 0.0243 0.1558 20 no phylogeny yes
## sigma^2.3 0.0215 0.1465 202 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 200) = 8200.4819, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 200) = 3.4784, p-val = 0.0636
##
## Model Results:
##
## estimate se tval df pval
## intrcpt 0.0391 0.0935 0.4179 18 0.6810
## scale(ramping, center = T, scale = F) 0.1532 0.0821 1.8650 200 0.0636
## ci.lb ci.ub
## intrcpt -0.1575 0.2356
## scale(ramping, center = T, scale = F) -0.0088 0.3151 .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1r2_ml(mod.persistent_ramping)## R2_marginal R2_conditional
## 0.03442667 0.56399312d.persistent_ramping <- filter(d, ramping != "NA" & brought_common_temp == "yes")
plot_continuous(d.persistent_ramping, mod.persistent_ramping, d.persistent_ramping$ramping,
"Heating rate (degrees/min)")
Risk of bias analysis
Analysis without imputed values
mod.imputed <- run.model(filter(d, imputed == "no"), ~1)
summary(mod.imputed)##
## Multivariate Meta-Analysis Model (k = 1000; method: REML)
##
## logLik Deviance AIC BIC AICc
## -81.0743 162.1486 170.1486 189.7756 170.1888
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0107 0.1034 133 no species_ID no
## sigma^2.2 0.0223 0.1492 133 no phylogeny yes
## sigma^2.3 0.0416 0.2041 1000 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 999) = 60216.9037, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.1873 0.0834 2.2447 132 0.0265 0.0222 0.3524 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.imputed, mod = "1", data = filter(d, imputed == "no"), group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.1873009 0.02224645 0.3523553 -0.3776772 0.752279my.orchard(mod.imputed, xlab = "dARR", mod = "1", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.07)
Analysis without concerning data
mod.concern <- run.model(filter(d, is_concern == "no"), ~1)
summary(mod.concern)##
## Multivariate Meta-Analysis Model (k = 852; method: REML)
##
## logLik Deviance AIC BIC AICc
## -48.5652 97.1305 105.1305 124.1161 105.1778
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0110 0.1047 126 no species_ID no
## sigma^2.2 0.0236 0.1536 126 no phylogeny yes
## sigma^2.3 0.0412 0.2031 852 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 851) = 54633.5157, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.1900 0.0866 2.1949 125 0.0300 0.0187 0.3614 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.concern, mod = "1", data = filter(d, is_concern == "no"), group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.1900224 0.01868273 0.3613621 -0.3811777 0.7612225my.orchard(mod.concern, xlab = "dARR", mod = "1", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.07)
Analyses without extreme values of dARR
Here, I performed we ran models after excluding values reaching arbitrary cutoffs: a) all values above 1.25 and below -1.25 b) All values under -0.5 c) All values under -0.15 (as in Gunderson & Stillman 2015) d) With all negative values taken as 0 (as in Morley et al. 2019)
While previous syntheses used such methods, it has been argued that negative responses can be biologically relevant (Terblanche & Hoffmann, 2020), and we believe it better captures the importance of measurement error, as well as potential “non-adaptive” responses ectotherms may express.
# Removing values above 1.25 and below -1.25
mod.extreme1.25 <- run.model(filter(d, dARR > -1.25 & dARR < 1.25), ~1)
summary(mod.extreme1.25)##
## Multivariate Meta-Analysis Model (k = 1079; method: REML)
##
## logLik Deviance AIC BIC AICc
## 11.4223 -22.8445 -14.8445 5.0869 -14.8072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0100 0.0999 138 no species_ID no
## sigma^2.2 0.0422 0.2054 138 no phylogeny yes
## sigma^2.3 0.0327 0.1809 1079 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 1078) = 58411.9685, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.1953 0.1129 1.7307 137 0.0858 -0.0278 0.4185 .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.extreme1.25, mod = "1", data = filter(d, dARR > -1.25 & dARR < 1.25),
group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.1953155 -0.02784668 0.4184777 -0.422581 0.813212my.orchard(mod.extreme1.25, xlab = "dARR", mod = "1", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.05)
## Removing values under -0.5
mod.extreme0.5 <- run.model(filter(d, dARR > -0.5), ~1) # Removed 34 effect sizes
summary(mod.extreme0.5)##
## Multivariate Meta-Analysis Model (k = 1055; method: REML)
##
## logLik Deviance AIC BIC AICc
## 45.8412 -91.6824 -83.6824 -63.8410 -83.6442
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0085 0.0920 137 no species_ID no
## sigma^2.2 0.1337 0.3656 137 no phylogeny yes
## sigma^2.3 0.0319 0.1786 1055 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 1054) = 57071.3506, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.2179 0.1980 1.1005 136 0.2731 -0.1737 0.6095
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.extreme0.5, mod = "1", data = filter(d, dARR > -0.5), group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.2178982 -0.1736688 0.6094652 -0.6953233 1.13112my.orchard(mod.extreme0.5, xlab = "dARR", mod = "1", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.05)
## Removing values under -0.15 (as in Gunderson & Stillman 2015)
mod.extreme0.15 <- run.model(filter(d, dARR > -0.15), ~1) # Removed 97 effect sizes
summary(mod.extreme0.15)##
## Multivariate Meta-Analysis Model (k = 992; method: REML)
##
## logLik Deviance AIC BIC AICc
## 126.9958 -253.9916 -245.9916 -226.3968 -245.9510
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0086 0.0928 136 no species_ID no
## sigma^2.2 0.1426 0.3776 136 no phylogeny yes
## sigma^2.3 0.0260 0.1611 992 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 991) = 48345.9519, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.2364 0.2043 1.1576 135 0.2491 -0.1675 0.6404
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.extreme0.15, mod = "1", data = filter(d, dARR > -0.15), group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.2364461 -0.1675137 0.6404059 -0.6887438 1.161636my.orchard(mod.extreme0.15, xlab = "dARR", mod = "1", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.05)
## Quantifying negative dARR values as 0 (as in Morley et al. 2019)
d <- d %>%
mutate(morley_dARR = ifelse(dARR < 0, 0, dARR)) # Convert negative values to 0
sum(with(d, morley_dARR == 0)) # 233 effect sizes concerned by this change## [1] 230mod.extreme0 <- rma.mv(yi = morley_dARR, V = VCV_dARR, mods = ~1, method = "REML",
test = "t", dfs = "contain", random = list(~1 | species_ID, ~1 | phylogeny, ~1 |
es_ID), R = list(phylogeny = phylo_matrix), data = d, sparse = TRUE)
summary(mod.extreme0)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## 128.2016 -256.4033 -248.4033 -228.4349 -248.3664
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0081 0.0902 138 no species_ID no
## sigma^2.2 0.0460 0.2144 138 no phylogeny yes
## sigma^2.3 0.0268 0.1637 1089 no es_ID no
##
## Test for Heterogeneity:
## Q(df = 1088) = 45685.9027, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## 0.2256 0.1173 1.9227 137 0.0566 -0.0064 0.4575 .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod_results(mod.extreme0, mod = "1", data = d, group = "species_ID")## name estimate lowerCL upperCL lowerPR upperPR
## 1 Intrcpt 0.2255555 -0.006423047 0.4575341 -0.3828002 0.8339113my.orchard(mod.extreme0, xlab = "dARR", mod = "1", alpha = 0.1, data = d, group = "species_ID",
whisker = 0.05)
Comparisons with the data from Morley et al. 2019. 10.1111/geb.12911
In our discussion, we discuss the overall levels of plasticity we report in developing ectotherms, and compare it with the results of previous syntheses. Here, we provide the code used to calculate the estimated ARR measured in Morley et al. 2019
d.morley <- read_csv("data/data_Morley_et_al_2019.csv")
# Quick processing
d.morley <- mutate(d.morley, habitat = ifelse(Habitat == "marine" | Habitat == "freshwater",
"aquatic", "terrestrial")) # merge freshwater and marine into 'aquatic'
d.morley <- filter(d.morley, life_history == "adult") # Only keep observations for adults for comparison
n_distinct(d.morley$Species) # 278 species## [1] 278length(d.morley$ctmax_ARR[d.morley$habitat == "aquatic"]) # 183 effect sizes aquatic## [1] 183length(d.morley$ctmax_ARR[d.morley$habitat == "terrestrial"]) # 153 effect sizes terrestrial## [1] 153library(lme4)
mod.morley <- lmerTest::lmer(ctmax_ARR ~ 1 + (1 | Species) + (1 | Source), data = d.morley)
summary(mod.morley)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ctmax_ARR ~ 1 + (1 | Species) + (1 | Source)
## Data: d.morley
##
## REML criterion at convergence: -135.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9270 -0.4131 -0.1081 0.2049 3.9181
##
## Random effects:
## Groups Name Variance Std.Dev.
## Species (Intercept) 0.02558 0.1599
## Source (Intercept) 0.00000 0.0000
## Residual 0.01636 0.1279
## Number of obs: 336, groups: Species, 278; Source, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.23544 0.01203 255.96993 19.57 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')# Remove observations classified as 'zero'
d.morley_nozero <- filter(d.morley, ctmax_ARR != "0") # Remove ARR values of zero; as they don't represent sampled data (all negative values were converted to 0)
mod.morley_nozero <- lmerTest::lmer(ctmax_ARR ~ 1 + (1 | Species) + (1 | Source),
data = d.morley_nozero)
summary(mod.morley_nozero)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ctmax_ARR ~ 1 + (1 | Species) + (1 | Source)
## Data: d.morley_nozero
##
## REML criterion at convergence: -146
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.8052 -0.5049 -0.2088 0.3835 4.2889
##
## Random effects:
## Groups Name Variance Std.Dev.
## Species (Intercept) 2.559e-02 0.1599738
## Source (Intercept) 3.397e-07 0.0005828
## Residual 1.319e-02 0.1148649
## Number of obs: 294, groups: Species, 245; Source, 2
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 0.27049 0.01239 35.03003 21.83 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.0246118 (tol = 0.002, component 1)Post-hoc analyses
Here, we present analyses that were suggested during peer-review. Note that we did not have a priori hypotheses and predictions for those analyses.
Body mass
Full dataset
d.mass <- filter(d, body_mass != "NA")
mod.body_mass <- run.model(d.mass, ~log(body_mass))
summary(mod.body_mass) # 494 effect sizes##
## Multivariate Meta-Analysis Model (k = 494; method: REML)
##
## logLik Deviance AIC BIC AICc
## -80.2948 160.5895 170.5895 191.5819 170.7130
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0093 0.0962 67 no species_ID no
## sigma^2.2 0.0013 0.0362 67 no phylogeny yes
## sigma^2.3 0.0562 0.2371 494 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 492) = 28344.9267, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 492) = 3.7683, p-val = 0.0528
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2144 0.0303 7.0661 65 <.0001 0.1538 0.2750 ***
## log(body_mass) 0.0122 0.0063 1.9412 492 0.0528 -0.0001 0.0246 .
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot_continuous(d.mass, mod.body_mass, log(d.mass$body_mass), "ln body mass (g)")
Direct effects
d.mass.direct <- filter(d.mass, brought_common_temp == "no")
mod.body_mass.direct <- run.model(d.mass.direct, ~log(body_mass))
summary(mod.body_mass.direct) # 411 effect sizes##
## Multivariate Meta-Analysis Model (k = 411; method: REML)
##
## logLik Deviance AIC BIC AICc
## -67.6426 135.2853 145.2853 165.3539 145.4342
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0035 0.0592 59 no species_ID no
## sigma^2.2 0.0022 0.0468 59 no phylogeny yes
## sigma^2.3 0.0597 0.2444 411 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 409) = 21512.7867, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 409) = 0.4708, p-val = 0.4930
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2577 0.0348 7.4076 57 <.0001 0.1880 0.3274 ***
## log(body_mass) 0.0043 0.0063 0.6861 409 0.4930 -0.0080 0.0167
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot_continuous(d.mass.direct, mod.body_mass.direct, log(d.mass.direct$body_mass),
"ln body mass (g)")
Persistent effects
d.mass.persistent <- filter(d.mass, brought_common_temp == "yes")
mod.body_mass.persistent <- run.model(d.mass.persistent, ~log(body_mass))
summary(mod.body_mass.persistent) # 83 effect sizes##
## Multivariate Meta-Analysis Model (k = 83; method: REML)
##
## logLik Deviance AIC BIC AICc
## 9.1862 -18.3724 -8.3724 3.5999 -7.5724
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0000 0.0000 11 no species_ID no
## sigma^2.2 0.0051 0.0714 11 no phylogeny yes
## sigma^2.3 0.0189 0.1373 83 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 81) = 608.7882, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 81) = 2.3102, p-val = 0.1324
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.0374 0.0546 0.6855 9 0.5103 -0.0861 0.1609
## log(body_mass) 0.0191 0.0126 1.5199 81 0.1324 -0.0059 0.0442
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot_continuous(d.mass.persistent, mod.body_mass.persistent, log(d.mass.persistent$body_mass),
"ln body mass (g)")
Age at sexual maturity
Full dataset
d.maturity <- filter(d, age_maturity != "NA")
mod.maturity <- run.model(d.maturity, ~log(age_maturity))
summary(mod.maturity) # 971 effect sizes##
## Multivariate Meta-Analysis Model (k = 971; method: REML)
##
## logLik Deviance AIC BIC AICc
## -55.0794 110.1588 120.1588 144.5401 120.2211
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0128 0.1132 114 no species_ID no
## sigma^2.2 0.0149 0.1220 114 no phylogeny yes
## sigma^2.3 0.0394 0.1984 971 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 969) = 53214.5206, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 112) = 0.6414, p-val = 0.4249
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.1937 0.0697 2.7790 112 0.0064 0.0556 0.3317 **
## log(age_maturity) 0.0112 0.0139 0.8009 112 0.4249 -0.0164 0.0387
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot_continuous(d.maturity, mod.maturity, log(d.maturity$age_maturity), "ln age at sexual maturity (years)")
Direct effects
d.maturity.direct <- filter(d.maturity, brought_common_temp == "no")
mod.maturity.direct <- run.model(d.maturity.direct, ~log(age_maturity))
summary(mod.maturity.direct) # 752 effect sizes##
## Multivariate Meta-Analysis Model (k = 752; method: REML)
##
## logLik Deviance AIC BIC AICc
## -53.4888 106.9777 116.9777 140.0781 117.0583
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0093 0.0964 98 no species_ID no
## sigma^2.2 0.0057 0.0757 98 no phylogeny yes
## sigma^2.3 0.0434 0.2083 752 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 750) = 34629.7412, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 96) = 0.0611, p-val = 0.8053
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.2223 0.0459 4.8451 96 <.0001 0.1313 0.3134 ***
## log(age_maturity) 0.0032 0.0131 0.2471 96 0.8053 -0.0227 0.0292
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot_continuous(d.maturity.direct, mod.maturity.direct, log(d.maturity.direct$age_maturity),
"ln age at sexual maturity (years)")
Persistent effects
d.maturity.persistent <- filter(d.maturity, brought_common_temp == "yes")
mod.maturity.persistent <- run.model(d.maturity.persistent, ~log(age_maturity))
summary(mod.maturity.persistent) # 219 effect sizes##
## Multivariate Meta-Analysis Model (k = 219; method: REML)
##
## logLik Deviance AIC BIC AICc
## 25.9138 -51.8275 -41.8275 -24.9280 -41.5432
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0007 0.0273 22 no species_ID no
## sigma^2.2 0.0226 0.1505 22 no phylogeny yes
## sigma^2.3 0.0220 0.1484 219 no es_ID no
##
## Test for Residual Heterogeneity:
## QE(df = 217) = 6342.1940, p-val < .0001
##
## Test of Moderators (coefficient 2):
## F(df1 = 1, df2 = 20) = 1.0701, p-val = 0.3133
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## intrcpt 0.0334 0.0896 0.3727 20 0.7133 -0.1536 0.2204
## log(age_maturity) -0.0198 0.0191 -1.0344 20 0.3133 -0.0596 0.0201
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot_continuous(d.maturity.persistent, mod.maturity.persistent, log(d.maturity.persistent$age_maturity),
"ln age at sexual maturity (years)")
Figures for publication
Note that each figure underwent cosmetic changes in Powerpoint.
Figure 2 (Design representation)
d_direct<-data.frame(category=c("A", "B", "C"),
count=c(700, 20, 146))
d_direct$fraction = d_direct$count / sum(d_direct$count)
d_direct$ymax = cumsum(d_direct$fraction)
d_direct$ymin = c(0, head(d_direct$ymax, n=-1))
ggplot(d_direct, aes(ymax=ymax, ymin=ymin, xmax=4, xmin=2.5, fill=category)) +
geom_rect(col="black") +
coord_polar(theta="y") + # Try to remove that to understand how the chart is built initially
xlim(c(1, 4))+theme_void()+scale_fill_manual(values=c("#7570B3", "#B370AC", "#CDCCE4"))+
theme(legend.background = element_rect(fill = "transparent"),
legend.box.background = element_rect(fill = "transparent"),
panel.background = element_rect(fill = "transparent"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
plot.background = element_rect(fill = "transparent", color = NA))
ggsave("output/direct_effect_sizes.png", height=12, width=12, dpi=800, bg="transparent")
d_persist<-data.frame(category=c("D", "E", "F"),
count=c(92, 76, 55))
d_persist$fraction = d_persist$count / sum(d_persist$count)
d_persist$ymax = cumsum(d_persist$fraction)
d_persist$ymin = c(0, head(d_persist$ymax, n=-1))
ggplot(d_persist, aes(ymax=ymax, ymin=ymin, xmax=4, xmin=2.5, fill=category)) +
geom_rect(col="black") +
coord_polar(theta="y") + # Try to remove that to understand how the chart is built initially
xlim(c(1, 4))+theme_void()+scale_fill_manual(values=c("#1B9E77", "#87EBCD", "#1B829E"))+
theme(legend.background = element_rect(fill = "transparent"),
legend.box.background = element_rect(fill = "transparent"),
panel.background = element_rect(fill = "transparent"),
panel.grid.major = element_blank(),
panel.grid.minor = element_blank(),
plot.background = element_rect(fill = "transparent", color = NA))
ggsave("output/persistent_effect_sizes.png", height=12, width=12, dpi=800, bg="transparent")Figure 3 (Phylogeny)
design_dat <- d.tree %>%
group_by(tip.label) %>%
summarise(direct = brought_common_temp == "no", persistent = brought_common_temp ==
"yes", CTmax = metric == "CTmax", LT50 = metric == "LT50") # Summarise, for each species, whether they tested direct effects, persistent effects, CTmax, or LT50
design_dat <- distinct(design_dat) # Only keep unique rows
design_dat$direct = as.numeric(design_dat$direct) # convert TRUE/FALSE to binary values
design_dat$persistent = as.numeric(design_dat$persistent) # convert TRUE/FALSE to binary values
design_dat$CTmax = as.numeric(design_dat$CTmax) #convert TRUE/FALSE to binary values
design_dat$LT50 = as.numeric(design_dat$LT50) #convert TRUE/FALSE to binary values
design_dat <- design_dat %>%
group_by(tip.label) %>%
summarise(direct = sum(direct), persistent = sum(persistent), CTmax = sum(CTmax),
LT50 = sum(LT50)) # calculate the sum for each species (i.e. if 1, the species has the given design or metric)
# In some cases, values were repeated, so need to replace '2' values by '1'
design_dat$direct[design_dat$direct == "2"] <- "1"
design_dat$persistent[design_dat$persistent == "2"] <- "1"
design_dat$CTmax[design_dat$CTmax == "2"] <- "1"
design_dat$direct = as.factor(design_dat$direct) # convert back to factor for the plot
design_dat$persistent = as.factor(design_dat$persistent)
design_dat$CTmax = as.factor(design_dat$CTmax)
design_dat$LT50 = as.factor(design_dat$LT50)
design_dat <- mutate(design_dat, direct_persistent = ifelse(direct == "1" & persistent ==
"1", "both", ifelse(direct == "1" & persistent == "0", "direct", "persistent"))) # If both direct and persistent, indicate 'both', if only direct, indicate 'direct', otherwise indicate 'persistent'
design_dat <- mutate(design_dat, metric_type = ifelse(CTmax == "1" & LT50 == "1",
"both", ifelse(CTmax == "1" & LT50 == "0", "CTmax", "LT50")))
d.tree_plot <- left_join(d.tree_plot, design_dat, by = "tip.label")
d.tree_plot$class <- gsub("Holuthuroidea", "Holothuroidea", d.tree_plot$class) # Fixed typo in the Class name
p <- ggtree(tree_for_plot, layout = "circular", lwd = 0.75)
p <- p %<+% d.tree_plot # link plot to data
p2 <- p + geom_fruit(geom = geom_tile, mapping = aes(fill = metric_type), width = 0.07,
offset = 0.085, col = "black", size = 0.3) + scale_fill_manual(values = c("gray70",
"white", "black"), name = "Metric") + theme(legend.direction = "vertical")
p3 <- p2 + new_scale_fill() + geom_fruit(geom = geom_tile, mapping = aes(fill = direct_persistent),
offset = 0.1, width = 0.07, col = "black", size = 0.3) + scale_fill_manual(values = c("#EAD295",
"#9994EA", "#95EA99"), name = "Design") + theme(legend.direction = "vertical")
d.tree_plot$n_es = as.numeric(d.tree_plot$n_es)
p4 <- p3 + new_scale_fill() + geom_fruit(geom = geom_bar, mapping = aes(x = n_es,
fill = class), stat = "identity", col = "black", orientation = "y", axis.params = list(axis = "x",
text.angle = -45, hjust = 0, text.size = 3), grid.params = list(alpha = 0.45),
offset = 0.085, pwidth = 0.55, alpha = 0.8, size = 0.3) + scale_fill_manual(values = c("#00BBDB",
"darkslategray4", "darkorange", "chartreuse", "indianred2", "lightsteelblue2",
"darkorchid", "darkseagreen1", "deeppink", "darkred", "forestgreen", "darkolivegreen2",
"#39568CFF", "darkgoldenrod2"), name = "Class") + theme(legend.position = "right",
legend.spacing.y = unit(5, "pt"), legend.box = "vertical", legend.box.just = "left",
legend.box.spacing = unit(-5, "pt"), legend.text = element_text(size = 12), legend.title = element_text(size = 14,
face = "bold"), legend.key.size = unit(20, "pt")) + guides(fill = guide_legend(order = 1)) # Display number of effect sizes
p4
# ggsave('output/phylo_tree_pub.svg', width=15, height=15, dpi=3000)Figure 4 (Meta-analytic mean)
p1 <- my.orchard(int_model, mod = "1", xlab = "dARR", alpha = 0.075, data = d, group = "species_ID",
whisker = 0.06) + scale_fill_manual(values = "#F3DF6B") + annotate("text", size = 6.5,
y = 1.75, x = 1.075, label = paste("0.190 [0.015, 0.364]"))
p2 <- my.orchard(mod.adjusted_mean, mod = "1", xlab = "dARR", alpha = 0.075, data = d,
group = "species_ID", whisker = 0.06) + scale_fill_manual(values = "#EB7B12") +
annotate("text", size = 6.5, y = 1.75, x = 1.075, label = paste("0.134 [0.002, 0.266]"))
p1/p2
# ggsave('output/meta-analysis_pub.svg', width=11, height=11, dpi=800)Figure 5 (Habitat variation)
d$habitat = as.factor(d$habitat)
d$habitat_reordered <- factor(d$habitat, c("terrestrial", "aquatic"))
mod.habitat_het_reordered <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~habitat_reordered -
1, method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~habitat_reordered | es_ID), struct = "HCS", rho = 0, R = list(phylogeny = phylo_matrix),
data = d, sparse = TRUE)
summary(mod.habitat_het_reordered)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -153.7523 307.5047 319.5047 349.4517 319.5824
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0948 138 no species_ID no
## sigma^2.2 0.0129 0.1136 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: habitat_reordered (nlvls = 2)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0082 0.0907 160 no terrestrial
## tau^2.2 0.0716 0.2677 929 no aquatic
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1087) = 56246.7341, p-val < .0001
##
## Test of Moderators (coefficients 1:2):
## F(df1 = 2, df2 = 136) = 8.5055, p-val = 0.0003
##
## Model Results:
##
## estimate se tval df pval ci.lb
## habitat_reorderedterrestrial 0.0595 0.0762 0.7814 136 0.4359 -0.0911
## habitat_reorderedaquatic 0.2087 0.0655 3.1856 136 0.0018 0.0792
## ci.ub
## habitat_reorderedterrestrial 0.2102
## habitat_reorderedaquatic 0.3383 **
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1my.orchard(mod.habitat_het_reordered, mod = "habitat_reordered", xlab = "dARR", alpha = 0.1,
data = d, group = "species_ID", whisker = 0.075) + scale_fill_manual(values = c("#996600",
"#33CCCC")) + annotate("text", size = 6.5, y = 1.75, x = 1.08, label = paste("0.051 [-0.091, 0.210]")) +
annotate("text", size = 6.5, y = 1.75, x = 2.08, label = paste("0.209 [0.079, 0.338]"))
# ('output/habitat_metareg_pub.svg', width=12, height=8, dpi=800)Figure 6 (Taxonomic variation)
# use the plural for taxonomic groups
d$taxonomic_group[d$taxonomic_group == "terrestrial invertebrate"] <- "Terrestrial invertebrates"
d$taxonomic_group[d$taxonomic_group == "reptile"] <- "Reptiles"
d$taxonomic_group[d$taxonomic_group == "amphibian"] <- "Amphibians"
d$taxonomic_group[d$taxonomic_group == "aquatic invertebrate"] <- "Aquatic invertebrates"
d$taxonomic_group[d$taxonomic_group == "fish"] <- "Fishes"
d$taxonomic_group = as.factor(d$taxonomic_group)
d$taxonomic_group_reordered <- factor(d$taxonomic_group, c("Terrestrial invertebrates",
"Reptiles", "Amphibians", "Aquatic invertebrates", "Fishes"))
mod.taxa_het_reordered <- rma.mv(yi = dARR, V = VCV_dARR, mods = ~taxonomic_group_reordered -
1, method = "REML", test = "t", dfs = "contain", random = list(~1 | species_ID,
~1 | phylogeny, ~taxonomic_group_reordered | es_ID), struct = "HCS", rho = 0,
R = list(phylogeny = phylo_matrix), data = d, sparse = TRUE)
summary(mod.taxa_het_reordered)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -54.5107 109.0214 133.0214 192.8824 133.3128
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0090 0.0946 138 no species_ID no
## sigma^2.2 0.0296 0.1722 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: taxonomic_group_reordered (nlvls = 5)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0069 0.0833 147 no Terrestrial invertebrates
## tau^2.2 0.0163 0.1278 27 no Reptiles
## tau^2.3 0.1878 0.4334 71 no Amphibians
## tau^2.4 0.1357 0.3684 221 no Aquatic invertebrates
## tau^2.5 0.0329 0.1813 623 no Fishes
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1084) = 56018.6724, p-val < .0001
##
## Test of Moderators (coefficients 1:5):
## F(df1 = 5, df2 = 133) = 1.9108, p-val = 0.0966
##
## Model Results:
##
## estimate se tval
## taxonomic_group_reorderedTerrestrial invertebrates 0.0488 0.1410 0.3464
## taxonomic_group_reorderedReptiles 0.0702 0.1733 0.4053
## taxonomic_group_reorderedAmphibians 0.1966 0.1761 1.1167
## taxonomic_group_reorderedAquatic invertebrates 0.1992 0.1286 1.5492
## taxonomic_group_reorderedFishes 0.2538 0.1265 2.0058
## df pval ci.lb
## taxonomic_group_reorderedTerrestrial invertebrates 133 0.7296 -0.2301
## taxonomic_group_reorderedReptiles 133 0.6859 -0.2726
## taxonomic_group_reorderedAmphibians 133 0.2661 -0.1517
## taxonomic_group_reorderedAquatic invertebrates 133 0.1237 -0.0551
## taxonomic_group_reorderedFishes 133 0.0469 0.0035
## ci.ub
## taxonomic_group_reorderedTerrestrial invertebrates 0.3278
## taxonomic_group_reorderedReptiles 0.4131
## taxonomic_group_reorderedAmphibians 0.5449
## taxonomic_group_reorderedAquatic invertebrates 0.4535
## taxonomic_group_reorderedFishes 0.5041 *
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1my.orchard(mod.taxa_het_reordered, mod = "taxonomic_group_reordered", xlab = "dARR",
alpha = 0.15, data = d, group = "species_ID", whisker = 0.075) + scale_fill_manual(values = c("#B39070",
"#8CBC4E", "#9AFB70", "#79D9F1", "#8DECE4")) + annotate("text", size = 6, y = 1.75,
x = 1.14, label = paste("0.049 [-0.230, 0.328]")) + annotate("text", size = 6,
y = 1.75, x = 2.14, label = paste("0.070 [-0.273, 0.413]")) + annotate("text",
size = 6, y = 1.75, x = 3.14, label = paste("0.197 [-0.152, 0.545]")) + annotate("text",
size = 6, y = 1.75, x = 4.14, label = paste("0.199 [-0.055, 0.454]")) + annotate("text",
size = 6, y = 1.75, x = 5.14, label = paste("0.254 [0.004, 0.504]"))
# ggsave('output/taxa_metareg_pub.svg', width=12, height=12, dpi=800)Figure 7 (Life history variation and persistence)
d$design_reordered<-factor(d$exp_design, c("F", "E", "D", "C", "B", "A"))
mod.design_het_reordered <- rma.mv(yi=dARR,
V=VCV_dARR,
mods= ~design_reordered-1,
method="REML",
test="t",
dfs="contain",
random=list(~1|species_ID,
~1|phylogeny,
~design_reordered|es_ID),
struct="HCS",
rho=0,
R = list(phylogeny = phylo_matrix),
data=d,
sparse=TRUE)
summary(mod.design_het_reordered)##
## Multivariate Meta-Analysis Model (k = 1089; method: REML)
##
## logLik Deviance AIC BIC AICc
## -122.9070 245.8140 273.8140 343.6388 274.2072
##
## Variance Components:
##
## estim sqrt nlvls fixed factor R
## sigma^2.1 0.0064 0.0798 138 no species_ID no
## sigma^2.2 0.0170 0.1306 138 no phylogeny yes
##
## outer factor: es_ID (nlvls = 1089)
## inner factor: design_reordered (nlvls = 6)
##
## estim sqrt k.lvl fixed level
## tau^2.1 0.0074 0.0859 55 no F
## tau^2.2 0.0134 0.1156 76 no E
## tau^2.3 0.0352 0.1876 92 no D
## tau^2.4 0.0155 0.1244 146 no C
## tau^2.5 0.3209 0.5665 20 no B
## tau^2.6 0.0753 0.2745 700 no A
## rho 0.0000 yes
##
## Test for Residual Heterogeneity:
## QE(df = 1083) = 47905.2266, p-val < .0001
##
## Test of Moderators (coefficients 1:6):
## F(df1 = 6, df2 = 1083) = 10.4670, p-val < .0001
##
## Model Results:
##
## estimate se tval df pval ci.lb ci.ub
## design_reorderedF 0.1015 0.0784 1.2956 1083 0.1954 -0.0522 0.2553
## design_reorderedE 0.1181 0.0825 1.4316 1083 0.1525 -0.0438 0.2800
## design_reorderedD -0.0818 0.0849 -0.9630 1083 0.3358 -0.2483 0.0848
## design_reorderedC 0.2502 0.0782 3.2004 1083 0.0014 0.0968 0.4036 **
## design_reorderedB 0.0981 0.1570 0.6249 1083 0.5321 -0.2099 0.4062
## design_reorderedA 0.2302 0.0740 3.1108 1083 0.0019 0.0850 0.3755 **
##
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1p1<-my.orchard(mod.design_het_reordered, mod="design_reordered", xlab="dARR", alpha=0.15, data=d, group="species_ID", whisker=0.1)+
scale_fill_manual(values=c("#1B829E","#87EBCD", "#1B9E77", "#CDCCE4", "#B370AC", "#9C98C9"))+
annotate("text", size=6, y=1.75, x=1.15, label=paste("0.102 [-0.052, 0.255]"))+ # F
annotate("text", size=6, y=1.75, x=2.15, label=paste("0.118 [-0.044, 0.280]"))+ # E
annotate("text", size=6, y=1.75, x=3.15, label=paste("-0.082 [-0.248, 0.085]"))+ # D
annotate("text", size=6, y=1.75, x=4.15, label=paste("0.250 [0.097, 0.404]"))+ # C
annotate("text", size=6, y=1.75, x=5.15, label=paste("0.098 [-0.210, 0.406]"))+ # B
annotate("text", size=6, y=1.75, x=6.15, label=paste("0.230 [0.085, 0.376]")) + # A
ylim(-2.5,3)
#ggsave("output/design_metareg_pub.png", width=12, height=12, dpi=800)
p2<-my.orchard(mod.persist_het, mod="brought_common_temp", xlab="dARR", alpha=0.1, data=d, group="species_ID", whisker=0.075)+
scale_fill_manual(values=c("#7570B3", "#1B9E77"))+
annotate("text", size=6, y=1.75, x=1.1, label=paste("0.224 [0.086, 0.363]"))+
annotate("text", size=6, y=1.75, x=2.1, label=paste("0.048 [-0.095, 0.192]"))+
ylim(-2.5, 3)
#ggsave("output/persist_metareg_pub.png", width=12, height=8, dpi=800)
p1/p2 + plot_layout(heights=c(12,8))
#ggsave("output/designs_persist.svg", width=14, height=20, dpi=1400)Software and packages versions
sessionInfo()## R version 4.1.3 (2022-03-10)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows 10 x64 (build 19042)
##
## Matrix products: default
##
## locale:
## [1] LC_COLLATE=English_Australia.1252 LC_CTYPE=English_Australia.1252
## [3] LC_MONETARY=English_Australia.1252 LC_NUMERIC=C
## [5] LC_TIME=English_Australia.1252
##
## attached base packages:
## [1] stats graphics grDevices utils datasets methods base
##
## other attached packages:
## [1] lme4_1.1-28 MuMIn_1.43.17 emmeans_1.7.2-9000003
## [4] ggExtra_0.9 here_1.0.1 ggstatsplot_0.8.0
## [7] ggdist_3.1.1 orchaRd_2.0 metaAidR_0.0.0.9000
## [10] metafor_3.0-2 Matrix_1.4-0 RColorBrewer_1.1-2
## [13] ggnewscale_0.4.5 tidytree_0.3.4 phytools_0.7-80
## [16] ggtreeExtra_1.3.3 ggtree_3.1.1.992 R.utils_2.10.1
## [19] R.oo_1.24.0 R.methodsS3_1.8.1 patchwork_1.1.1
## [22] ape_5.5 rotl_3.0.11 maps_3.3.0
## [25] viridis_0.6.1 viridisLite_0.4.0 DataExplorer_0.8.2
## [28] kableExtra_1.3.4 forcats_0.5.1 stringr_1.4.0
## [31] dplyr_1.0.7 purrr_0.3.4 readr_2.1.1
## [34] tidyr_1.1.3 tibble_3.1.3 ggplot2_3.3.5
## [37] tidyverse_1.3.1
##
## loaded via a namespace (and not attached):
## [1] utf8_1.2.2 tidyselect_1.1.1
## [3] htmlwidgets_1.5.3 grid_4.1.3
## [5] combinat_0.0-8 gmp_0.6-2
## [7] munsell_0.5.0 codetools_0.2-18
## [9] effectsize_0.4.5 rentrez_1.2.3
## [11] miniUI_0.1.1.1 withr_2.4.3
## [13] colorspace_2.0-2 highr_0.9
## [15] knitr_1.37 rstudioapi_0.13
## [17] stats4_4.1.3 ipmisc_6.0.2
## [19] ggsignif_0.6.2 labeling_0.4.2
## [21] mnormt_2.0.2 farver_2.1.0
## [23] datawizard_0.2.0.1 rprojroot_2.0.2
## [25] coda_0.19-4 vctrs_0.3.8
## [27] treeio_1.16.1 generics_0.1.1
## [29] TH.data_1.0-10 clusterGeneration_1.3.7
## [31] xfun_0.29 BWStest_0.2.2
## [33] R6_2.5.1 ggbeeswarm_0.6.0
## [35] rmdformats_1.0.2 BayesFactor_0.9.12-4.2
## [37] cachem_1.0.5 reshape_0.8.8
## [39] assertthat_0.2.1 promises_1.2.0.1
## [41] networkD3_0.4 scales_1.1.1
## [43] multcomp_1.4-17 beeswarm_0.4.0
## [45] gtable_0.3.0 multcompView_0.1-8
## [47] phangorn_2.7.1 sandwich_3.0-1
## [49] rlang_0.4.11 MatrixModels_0.5-0
## [51] zeallot_0.1.0 systemfonts_1.0.2
## [53] scatterplot3d_0.3-41 PMCMRplus_1.9.0
## [55] splines_4.1.3 lazyeval_0.2.2
## [57] broom_0.7.12 BiocManager_1.30.16
## [59] yaml_2.2.1 modelr_0.1.8
## [61] backports_1.4.1 httpuv_1.6.1
## [63] tools_4.1.3 bookdown_0.22
## [65] ellipsis_0.3.2 jquerylib_0.1.4
## [67] WRS2_1.1-3 Rcpp_1.0.7
## [69] plyr_1.8.6 progress_1.2.2
## [71] prettyunits_1.1.1 pbapply_1.5-0
## [73] correlation_0.7.0 zoo_1.8-9
## [75] haven_2.4.3 ggrepel_0.9.1
## [77] fs_1.5.0 magrittr_2.0.1
## [79] data.table_1.14.2 reprex_2.0.1
## [81] tmvnsim_1.0-2 mvtnorm_1.1-3
## [83] mime_0.12 hms_1.1.1
## [85] evaluate_0.14 xtable_1.8-4
## [87] XML_3.99-0.6 pairwiseComparisons_3.1.6
## [89] readxl_1.3.1 gridExtra_2.3
## [91] compiler_4.1.3 crayon_1.4.2
## [93] minqa_1.2.4 htmltools_0.5.1.1
## [95] later_1.2.0 mc2d_0.1-21
## [97] tzdb_0.2.0 aplot_0.0.6
## [99] expm_0.999-6 lubridate_1.8.0
## [101] DBI_1.1.2 SuppDists_1.1-9.5
## [103] kSamples_1.2-9 formatR_1.11
## [105] dbplyr_2.1.1 MASS_7.3-55
## [107] boot_1.3-28 cli_3.0.1
## [109] quadprog_1.5-8 parallel_4.1.3
## [111] insight_0.14.2 igraph_1.2.6
## [113] pkgconfig_2.0.3 rncl_0.8.4
## [115] rvcheck_0.1.8 statsExpressions_1.1.0
## [117] numDeriv_2016.8-1.1 xml2_1.3.3
## [119] paletteer_1.4.0 svglite_2.0.0
## [121] vipor_0.4.5 bslib_0.2.5.1
## [123] webshot_0.5.2 estimability_1.3
## [125] rvest_1.0.2 distributional_0.3.0
## [127] digest_0.6.27 parameters_0.14.0
## [129] rmarkdown_2.11 cellranger_1.1.0
## [131] fastmatch_1.1-3 shiny_1.6.0
## [133] gtools_3.9.2 nloptr_1.2.2.2
## [135] lifecycle_1.0.1 nlme_3.1-155
## [137] jsonlite_1.7.2 fansi_0.5.0
## [139] pillar_1.6.5 lattice_0.20-45
## [141] fastmap_1.1.0 httr_1.4.2
## [143] plotrix_3.8-1 survival_3.2-13
## [145] glue_1.4.2 bayestestR_0.10.5
## [147] stringi_1.7.6 sass_0.4.0
## [149] performance_0.7.3 rematch2_2.1.2
## [151] memoise_2.0.0 mathjaxr_1.4-0
## [153] Rmpfr_0.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